Hdevelop Users Guide
Hdevelop Users Guide
This manual is a guide to HDevelop, the interactive development environment for the HALCON machine vision
library. It is intended for users who want to use HDevelop as a convenient gateway to the HALCON library or
who want to deploy and test machine vision applications with it. It is not an introduction to the HALCON machine
vision library. A working knowledge of the concepts of HALCON is assumed. Please refer to the Quick Guide to
become acquainted with HALCON. In-depth knowledge of image processing is not required to start working with
HDevelop.
The manual is divided into the following chapters:
• Introducing HDevelop
This chapter explains the basic concepts of HDevelop.
• Getting Started
This chapter explains how to start HDevelop. It provides a quick overview of the graphical user interface,
and shows you how to run the supplied example programs.
• Acquiring Images with HDevelop
This chapter explains the fundamental part of machine vision applications – how to acquire images.
• Programming HDevelop
This chapter explains how to develop applications in HDevelop.
• HDevelop Procedures
This chapter introduces the concepts of breaking a large program into small maintainable and reusable units.
• Graphical User Interface
This chapter explains the graphical user interface of HDevelop and how to interact with it.
• HDevelop Assistants
This chapter describes how to use the machine vision assistants of HDevelop.
• HDevelop Language
This chapter explains the syntax and semantics of the language used in HDevelop expressions.
• Remote Debugging
This chapter explains how remote debugging is supported.
• Code Export
This chapter explains the export of an HDevelop program to C, C++, Visual Basic.NET, or C#.
• Appendix
The appendix contains a glossary, a list of predefined color names, a list for command line usage, and a list
of keyboard shortcuts.
Symbols
The following symbols are used within the manual:
1 Introducing HDevelop 11
1.1 Facts About HDevelop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2 HDevelop XL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3 Terminology & Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2 Getting Started 15
2.1 Running HDevelop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Start Dialog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3 User Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4 Organizing HDevelop’s Workspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Running Example Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5 HDevelop Procedures 43
5.1 Procedure Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.2 File Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.2.1 File Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.2.2 HDevelop Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.2.3 Procedure Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.2.4 Libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.3 Procedure Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.4 Procedure Locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.5 Procedure Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.6 Protected Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.7 Procedure Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.8 Legacy Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.9 Just-in-Time Compilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
A Glossary 319
Index 335
Introducing HDevelop 11
Introduction
Chapter 1
Introducing HDevelop
HDevelop is a tool box for building machine vision applications. It facilitates rapid prototyping by offering a
highly interactive programming environment for developing and testing machine vision applications. Based on
the HALCON library, it is a versatile machine vision package suitable for product development, research, and
education.
There are four basic ways to develop image analysis applications using HDevelop:
• Rapid prototyping in the interactive environment HDevelop.
You can use HDevelop to find the optimal operators or parameters to solve your image analysis task, and then
build the application using various programming languages, for example, C, C++, C#, or Visual Basic.NET.
• Development of an application that runs within HDevelop.
Using HDevelop, you can also develop a complete image analysis application and run it within the HDevelop
environment. The example programs supplied with HDevelop can be used as building blocks for your own
applications.
• Execution of HDevelop programs or procedures using HDevEngine.
You can directly execute HDevelop programs or procedures from an application written in C++ or any lan-
guage that can integrate .NET objects using HDevEngine. This is described in detail in the Programmer’s
Guide, part VI on page 139.
• Export of an application as C, C++, Visual Basic.NET, or C# source code.
Finally, you can export an application developed in HDevelop as C, C++, Visual Basic.NET, or C# source
code. This program can then be compiled and linked with the HALCON library so that it runs as a stand-
alone (console) application. Of course, you can also extend the generated code or integrate it into existing
software.
Let’s start with some facts describing the main characteristics of HDevelop.
2 HDevelop immediately displays the results of operations. You can try different operators and/or parame-
ters, and immediately see the effect on the screen. Moreover, you can preview the results of an operator
without changing the program.
2 Several graphical tools allow you to examine iconic and control data online. For example, you can extract
shape and gray value features by simply clicking onto the objects in the graphics window, or inspect the
histogram of an image interactively and apply real-time segmentation to select parameters.
2 Built-in graphical assistants provide interactive interfaces to more complex machine vision tasks. The
assistants can also generate HDevelop code in the current program.
2 Variables with an automatic garbage collection are used to manage iconic objects or control values.
2 HDevelop supports just-in-time compilation of procedures for optimized performance as well as calling
procedures as separate sub-threads.
1.2 HDevelop XL
In addition to the standard HDevelop, there is also a variant called HDevelop XL, which is based on HALCON
XL. The user interface is identical, but underneath HALCON XL is optimized for large images. In the remainder
of this document, when we refer to HDevelop you can substitute HDevelop XL if that is the variant you will be
using.
HDevelop adheres to well-established conventions and usage patterns regarding its graphical user interface. Most
of the terminology explained here will have become second nature to most users and may most likely be skimmed
over.
Mouse Usage
click A single click with the left mouse button,
for example, to mark and select items or to activate buttons. To
select multiple items, hold down the Ctrl key and click the desired items. To select many items from a list,
click the first item, hold down the Shift key and click the last item. All intermediate items are then also
selected.
double-click Two quick successive clicks with the left mouse button, for example, to open dialogs of selected
items. Double-clicks are mostly shortcuts for single clicks followed by an additional action.
right-click A single click with the right mouse button to access additional functionality of the user interface, for
example, context-sensitive menus. Clicking the right mouse button also ends interactive drawing functions
in HDevelop.
drag Keeping the left mouse button pressed while moving the mouse and finally releasing the mouse button.
Typically used to move items, resize windows, select multiple items at once, for example, program lines, or
to draw shapes.
drag-and-drop HDevelop supports drag-and-drop of image files, dictionaries, 3D object models, DL models, and
HDevelop programs from other applications. You can, for example, drag an HDevelop program from a file
browser and drop it on HDevelop to load it.
middle mouse button With three-button mice, the middle mouse button is used under Linux to paste text from
the clipboard into text fields.
mouse wheel Most recent three-button mice combine the middle mouse button with a scrolling wheel. HDevelop
supports the mouse wheel in many places. The mouse wheel operates the GUI element under the mouse
cursor. Using the mouse wheel you can, for instance, quickly scroll large program listings, select values
from lists or perform continuous zooming of displayed images. In general, windows that provide a scroll
bar can be quickly scrolled with the mouse wheel. Furthermore, the values of spinner boxes (text fields that
expect numerical data) can be decremented and incremented with the mouse wheel.
1.3 Terminology & Usage 13
Keyboard Usage
HDevelop is very keyboard-friendly. Most functions of the graphical user interface that can be operated using the
Introduction
mouse can be accessed from the keyboard as well. Many of the most important functions are available through
keyboard shortcuts, which are worthwhile memorizing. When programming with HDevelop, keeping both hands
on the keyboard can increase the productivity. Therefore, many navigational tasks like selecting parameter fields
or selecting values from lists can easily be done using just the keyboard. The most common keyboard functions
are listed in the appendix D on page 327.
To make it easier for you to memorize the keyboard shortcuts, many of them are introduced by a common com-
bination to
indicate the context. For example, many shortcuts related to the graphics window are introduced by
pressing Ctrl+Shift+G followed by another key, for example, Ctrl+Shift+G,Del clears the graphics window. Because it
is often easier to keep Ctrl+Shift pressed when hitting the second key the alternative Ctrl+Shift+G,Ctrl+Shift+Del is also
allowed.
Certain key combinations
may conflict with keyboard functions of the operating system or the window manager.
For example, using Ctrl+Alt+Cursor Keys in the graphics window pans the displayed image while displaying pixel
information. Under Windows it may also change the screen orientation. See your system documentation on how
to disable or change the conflicting key bindings in this case.
Abbreviations
BP breakpoint
IC insert cursor
GUI graphical user interface
PC program counter
XLD extended line description (see also chapter A on page 319)
14 Introducing HDevelop
Getting Started 15
Chapter 2
Getting Started
Getting Started
In the following it is assumed that HALCON has already been installed as described in the Installation Guide.
Windows
Under Windows, HDevelop is usually started from the Start menu.
You can also start HDevelop from the Windows command prompt or from the Run app (for example, press
Windows logo key + R ). This makes it easy to pass optional command line switches:
hdevelop
Linux
Under Linux, HDevelop is started from the shell:
hdevelop &
16 Getting Started
Introductory material
Documentation
For a first introduction to HDevelop, try the links listed under “Getting Started”. To close the start dialog without
any further action, press Esc . If you have accidentally closed the start dialog, it can be re-opened from HDevelop’s
Help menu.
When HDevelop is started for the first time it looks similar to figure 2.2. The following windows are available by
default:
2.4 Organizing HDevelop’s Workspace 17
Variable window
Program variables can be watched in this window. It displays all variables of the current procedure and their
current values. Iconic variables are displayed as thumbnails, whereas control variables are displayed as text.
The layout of this window can be switched between horizontal and vertical splitting by double-clicking the
separator. See also Variable Window (page 160).
You can double-click iconic variables to display them in the active graphics window. Double-clicking control
variables opens an inspection window with a formatted list of the current values and statistical data. See also
Inspecting Variables (page 167).
Getting Started
Canvas window with graphics window
The canvas window is a container for graphics windows. It contains one empty graphics window, at first
startup of HDevelop. Graphics windows are used to display iconic data like images, regions, and XLDs, for
more information see also Graphics Window (page 73). Each graphics window provides its own toolbar to
adapt the image display to your needs and to, for example choose between different zoom steps. By default,
the toolbar of graphics windows on the canvas is hidden. To display it, right-click in the canvas window and
activate Show Graphics Window Toolbar.
The canvas window helps you to organize multiple graphics windows. You can arrange and resize the
graphics windows on the canvas freely. For this, right-click in the canvas and activate Show Graphics
Window Frame. Now you can drag the title bar of the graphics window to move it or drag the bottom right
corner to resize the window. You can also detach the graphics windows from the canvas to display them
floating. See also section 6.4 on page 66.
Menu
Tool bar
Program Window
This window displays the current program. It provides syntax highlighting with user-definable colors. The
left column displays the program line numbers. The small black triangle is the insert cursor, which is where
new program lines will be added. In the following, it is referred to as IC. The green arrow is the program
counter which marks the next line to be executed. In the following, the program counter is referred to as PC.
You can also add or remove breakpoints in the current program in this column. These will halt the program
execution at user-defined places so that intermediate results may be examined. See also Program Window
(page 127).
When adding new lines or modifying existing lines, advanced autocompletion (page 130) features speed up
typing and help keeping the program consistent. Program lines can also be modified by double-clicking
them and editing them in the operator window.
Status bar
HDevelop offers various possibilities to organize and personalize your workspace. The following paragraphs show
which changes to HDevelop’s layout are possible.
part. The indicator in the middle tabs the window. The window joins the already docked window as a tab
card.1
i 2. When dragging the window and holding down the Ctrl key, the drop indicators disappear.
Alternatively, click Dock window in the top right corner of the window. The window docks or tabs at its prior
position.
To undock a window:
• Drag the window at the title bar/tab card and drop it apart from the drop indicators. The window will keep
its current size.
• Or click Undock window in the top right corner of the window or double-click the title bar of the window.
The window will resize to its prior size as floating window.
The position and size of docked windows is restored after closing and reopening HDevelop. There are only a few
exceptions to this rule: HDevelop Assistants (Calibration, Measure...), Call Stack, Find Replace Dialog, Output
Console Preferences, ROI Window, and Inspection Windows (ROI Inspect, Handle Inspect, Variable Inspect...).
Note that the main HDevelop windows, namely the graphics window, the program window, and the variable win-
dow are always restored, even if they have been closed before.
Auto-hide: Hide the docked window in one of the 4 side bars (top, left, right, bottom). Click the docked
window’s entry on the side bar to show it again. It will be displayed overlaid (on top of the current layout).
Disable auto-hide: Restore the docked window into the layout.
HALCON comes with many HDevelop example programs from a variety of application areas. These range from
simple programs that demonstrate a single aspect of HALCON or HDevelop to complete machine vision solutions.
As an introduction to HDevelop, we recommend trying some of these programs to quickly get accustomed to the
way HDevelop works.
The example program “Explore the Power of HALCON” demonstrates many different capabilities of HALCON in
one program. It can be started from the start dialog, see figure 2.1 on page 16 (2). Running this program is highly
recommended to get a good overview of the many application areas of HALCON.
Examples
“Explore the Power of HALCON” starts up automatically when loaded from the start dialog. After loading it
manually or loading one of the other example programs click or press F5 to start it, see figure 2.3 (1).
For more information on how to run programs with a floating canvas or graphics window, see section 6.4.1 on page
66.
The example programs have been categorized by application area, industry, method, and operator usage. A special
category “New in version” groups examples by their appearance in specific HALCON releases. Browsing these
categories, you can quickly find example programs that cover image processing problems that you may wish to
solve with HALCON. These programs may serve as a foundation for your own development projects.
To open the example program browser, click File . Browse HDevelop Example Programs...
For more information, see section 6.5 on page 69. Click to run the program.
20 Getting Started
Figure 2.4: The variable and program window after loading the example program.
PC (program counter)
Current procedure
Iconic variables
Control variables
The program lines of the loaded example program are now displayed in the program window. The PC is set to
the first executable line of the program (leading comments are ignored). The variable window is also updated: It
lists the variables that are used in the main procedure, which is initially the current procedure. The variables are
currently uninstantiated, meaning their current value is undefined. This is indicated by the question mark (?). Both
windows are displayed in figure 2.4.
Run continuously
Run step-by-step
Stop
Reset
2.5 Running Example Programs 21
Or, you can convert HDevelop programs to other programming languages without opening the graphical user
Examples
interface at all. A full list of the supported command line switches is available with the following command:
hdevelop --help
See appendix C.1 on page 323 for a listing of the available switches, and some example uses of the command line.
22 Getting Started
Acquiring Images With HDevelop 23
Chapter 3
Image Acquisition
Image acquisition is crucial for machine vision applications. It will usually be an early if not the first step in your
programming projects. This chapter explores the different ways of image acquisition in HDevelop.
Drag-and-Drop
The easiest way to read an image is to simply drag it from a file browser to the HDevelop window and drop it there.
When the file is dropped, HDevelop opens the dialog Read Image (see figure 3.1).
This dialog displays the full path of the image and automatically proposes a variable name derived from the file
name. This name can be edited, or another iconic variable name from the current program can be selected from
the drop-down list.
Furthermore, a preview of the image and basic image properties are displayed in the dialog (width, height, color
type, and number of channels). If you picked the wrong image, you can select another one from the same directory
by pressing the button next to the file name. This will open a file browser native to the operating system. On
Windows, you may be able to switch to thumbnail view in this dialog. When another image is selected, the dialog
is updated accordingly.
24 Acquiring Images With HDevelop
When you click the button OK, the instruction read_image is added to the current program. With the setting of
Insert Position you determine where the instruction will be put: At the IC or the PC. If you changed your
mind about reading the selected image at all, click Cancel.
File Name You can enter the name of an image file into this field. A relative file name is looked up in several
directories in that order:
1. The current working directory (., a single dot) is the directory HDevelop was started from.
2. The subdirectory images of the directory specified by the environment variable HALCONROOT.
3. The directories specified by the environment variable HALCONIMAGES.
4. The directories used in previous invocations of Read Image....
The first matching image file is displayed as a thumbnail preview along with its width, height, color type and
number of channels.
Alternatively, open a file selection dialog by selecting a predefined directory from the File Name combo
box or clicking the browse button. The latter will start browsing in the current working directory, or in the
last used directory. Depending on the operating system you may be able to switch to a thumbnail view in the
file selection dialog.
Variable Name HDevelop suggests a variable name derived from the selected file name. You may adopt or edit
this name. If you want to use a name of an already created iconic variable, this combo box offers you all
known iconic variable names. Simply click the arrow on the right side of the combo box to select a variable
name. Note that the reuse of a variable deletes the old content and replaces it with the new image.
After selecting a file name, click OK to load the image into HDevelop. The operator read_image is inserted at the
specified insert position (IC or PC). The specified iconic variable is updated in the variable window and the image
is displayed in the active graphics window. Clicking Cancel aborts the operation.
By default, an absolute path to the selected image is inserted. You can instruct HDevelop to use relative path names
(see General Options . General Options in the preferences).
When images are read as described above, they are automatically displayed in the active graphics window. This
is the default behavior, but the automatic display of images can be suppressed if desired, for example, to speed up
computationally intensive programs.
Initially, new images are fitted into the graphics window at a 1:1 aspect ratio with unused bars on the left and right
side. The aspect ratio will be preserved when resizing the graphics window. Using the tool bar of the graphics
window (page 75) you can easily zoom the image or change the resize mode from the default Keep Aspect
Ratio behavior. Furthermore, you can change the default resize behavior under Edit . Preferences . General
Options. This default resize behavior applies to graphics windows that are opened via menu, Visialization .
Open Graphics Window... or Window . Open Graphics Window.
An iconic view of the loaded image is also displayed in the variable window. When the image is cleared in the
graphics window, the image can always be restored by double-clicking the corresponding icon in the variable
window. See also section Displaying Iconic Variables (page 163).
3.3 Image Acquisition Assistant 25
The image acquisition assistant is a powerful tool to acquire images from files (including AVI files), directories
or image acquisition devices supported by HALCON through image acquisition interfaces. To use this assistant,
select Assistants . Open New Image Acquisition. The window is displayed in figure 3.2. It features several
tab cards that can be stepped through one after another. Ultimately, the assistant generates HDevelop code that can
be inserted into the current program. Select the entry Help in the menu of the image acquisition assistant to open
its online help.
Image Acquisition
Figure 3.2: Image acquisition assistant.
The tab card Source determines the acquisition method and the image source. In the default setting images are
acquired from files. This is described in the following section. Alternatively, images are acquired from an image
acquisition device, for example, a camera. This is described in section 3.3.2 on page 27.
You can specify a selection of image files or also drag and drop a complete directory to load images from. Make
sure the radio button Image File(s) is selected in the tab card Source. You can directly enter image names or
the name of a directory into the text field. Multiple image names are separated by a semicolon. Usually, it is more
convenient to use one of the following buttons:
View Images
You can single-step through the selected images by clicking the Snap button (see figure 3.3). Each time you click
the button, the next image is displayed in the active graphics window. You can also loop through the images by
clicking the button Live. This is especially useful for animations. Both functions are also available from the menu
Calibration.
26 Acquiring Images With HDevelop
Connect
Snap (single-step images)
Generate Code
Switch to the tab card Code Generation, and specify a variable name in the text field Image Object. You can
later access the image in the program by this name. If multiple images or a directory were selected in the tab
card Source, the image acquisition assistant will read the images in a loop. In this case the following additional
variable names need to be specified:
Loop Counter: The name of the loop index variable. While looping over the images in the program, this variable
will contain the object number of the current image.
Image Files: The name of the variable that will contain the names of the selected images.
Click Code Preview to inspect the code that would be generated from the currently specified parameters.
Click Insert Code to generate the code and insert it at the position of the IC in the current program.
The following piece of code is an example generated from three selected images. It is a self-contained HDevelop
program that runs without alteration.
Select Image Acquisition Interface in the Source tab. The drop-down list below the radio button becomes
active. Initially, it lists all image acquisition interfaces supported by HALCON. You can tidy this list by clicking
the button Auto-detect Interfaces . HDevelop will then probe all the image acquisition interfaces and
remove those that do not respond. Probing the interfaces might cause the system to hang due to wrongly installed
drivers or hardware failures. If there are unsaved changes in the current program, HDevelop will display a warning
dialog. You are advised to save the changes before you proceed. You can also ignore the warning and proceed, or
cancel the process. After the interfaces have been probed, you can select the desired image acquisition interface
from the list .
Selecting the entry Help from the menu of the image acquisition assistant will open the online help for the selected
image acquisition interface.
Image Acquisition
Figure 3.5: Source selection (example).
Auto-detect Interfaces
Interface list
Range of parameters
Update Image
Refresh
Depending on the parameter type, different selection methods are enabled. As an example, parameters with a
defined range of values can be specified by dragging a slider or entering the value parametrically. If a value is
changed, a reset button to the right is activated . Some parameters provide a check box which attempts to set the
parameter automatically if clicked.
3.3.2 Acquiring Images Through Image Acquisition Interfaces 29
If Update Image is checked, parameter changes are immediately reflected in the graphics window by acquiring
a new image. The button Refresh updates the list of parameters, which is useful if parameters have side effects.
You can reset all parameters to their default values at once by clicking Reset All.
EasyParams
The tab card EasyParams contains a preselection of the most important camera parameters. These parameters are,
in most cases, enough to rapidly obtain an acceptable image in the Image Acquisition Assistant. Because
the EasyParams rely on the GenICam features of your device, the tab card will only be available for GenICam-
based image acquisition interfaces, and after connecting to a device.
Each one of the EasyParams internally controls multiple GenICam features of the connected device. See figure 3.8
for an example parameter list. Refer to Image Acquisition Interface Reference for detailed information. You can
click the help button of the assistant to get to the corresponding page automatically.
Image Acquisition
Figure 3.8: EasyParams.
Generate Code
The tab card Code Generation turns the settings made in the other tab cards into executable code. The basic
structure of the code and the corresponding variable names can be specified.
Control Flow
Initialization Only: Generate only code to initialize the image acquisition interface with the parameters
specified in the other tab cards and to close it down properly. Additional code for image acquisition and processing
can be added later.
Acquire Single Image: Also generate code to acquire an image.
Acquire Images in Loop: Also add a loop around the image acquisition code. Further image processing can
be added inside this loop.
The image acquisition interface is addressed by a so-called handle. The variable name of this handle can be
specified in the text field Connection Handle. The variable name of the acquired image(s) can be set in Image
Object.
Click Code Preview to inspect the code. Click Insert Code to generate the code in the program window at the
IC.
Here is a code example:
while (true)
grab_image_async (Image, AcqHandle, -1)
* Image Acquisition 01: Do something
endwhile
close_framegrabber (AcqHandle)
After the generated code has been inserted into the program window, HDevelop internally keeps the code linked
to the corresponding assistant. This link is kept until the assistant is quit using the menu entry File . Exit
Assistant. If you close the assistant by using the menu entry File . Close Dialog or the close icon of the
window, the assistant can be restored from the top of the menu Assistants.
You can change the settings inside the assistant and update the generated code accordingly. The code preview will
show you exactly how the generated code lines will be updated. Furthermore, you can delete the generated code
lines, or release them. When code lines are released, the internal link between the assistant and those lines is cut
off. Afterwards, the same assistant can generate additional code at a different place in the current program.
Programming With HDevelop 31
Chapter 4
This chapter explains how to use HDevelop to develop your own machine vision applications. Preferably, actively
follow the examples in a running instance of HDevelop. In the following, it is assumed that the preferences of
HDevelop are set to the default values. This is always the case after a fresh installation of HALCON. If you are
uncertain about the current settings, you can always start HDevelop with the default settings by invoking it from
the command line in the following way (see also chapter 2 on page 15):
hdevelop -reset_preferences
Programming
This chapter deals with a simple example. Given is the image displayed in figure 4.1. The objective is to count the
clips and determine their orientation.
Start HDevelop or, if it is still running, click File . New Program to start a new program. HDevelop will notify
you if there are unsaved changes in the current program. If it does, click Discard to throw away the changes and
start anew.
The first thing to do is read the image and store it in an iconic variable. From chapter 3 on page 23 we already
know that we can simply drag an image to the HDevelop window. We also know that this inserts the operator
read_image into the program. Therefore, we can just as well insert the operator directly.
32 Programming With HDevelop
Figure 4.2: Matching operators after typing read_ and pressing Return.
After selecting an operator, its parameters are displayed in the operator window. They are grouped by iconic and
control parameters. The icons next to the parameter names denote the parameter type: Input vs. output (see
figure 4.3). The semantic type is displayed to the right of the parameters. Parameters are specified in the text fields.
The first parameter gets the input focus.
Enter Clip into the text field Image.
The image will be stored in this variable.
Next, enter ’clip’ into the text
field FileName. You can press Tab to go to the next input field. Pressing Shift+Tab takes you back to the previous
field. This way you can enter all parameters without using the mouse.
Click OK to add the operator to the current program and execute it. This will do the following:
Programming
Semantic type
Data type
• Set MinGray and MaxGray to 0 and 30, respectively. This will select the dark pixels in the image.
• Click Apply. This button executes the operator without adding it to the program. Additionally, it keeps
the current parameters open for editing. This way, different settings can be tried and the result can be seen
immediately. The selected pixels are stored in the output variable Region, which is displayed in the variable
window. The region is an image mask: White pixels are selected while black pixels are not.
• The region is also displayed as an overlay in the graphics window. The selected pixels are displayed in red
by default.
• Click Enter to add the operator to the program window. Contrary to clicking OK, this does not execute the
operator. Note that the variable Region keeps its value but is no longer displayed in the graphics window.
Also, the PC is not advanced, indicating that the second line of the program is yet to be executed.
• Adding program lines with Enter is especially useful if some input parameters use variable names that will
be added to the program at a later time.
Successor
• Click the just inserted program line to select it. You can let HDevelop suggest operators based on the selected
line.
• Open the menu Suggestions . Successors. This menu is filled dynamically to show typical successors
of the currently selected operator. We want to split the selected pixels into contiguous regions.
• Move the mouse pointer over the menu entries. The status bar displays a short description of the highlighted
operator. Any operator selected through this menu is transferred to the operator window.
• Click the operator connection.
• By clicking OK. Two program lines are executed: The threshold operation and the connection operation.
Clicking OK executes from the PC to the IC.
• In the graphics window, the contiguous regions calculated by the operator connection are now displayed
in alternating colors.
After having executed the three lines of the program, the graphics window displays three layers of iconic variables:
4.7 Inspecting Variables 35
Programming
Figure 4.5: Changing the display parameters.
You cannot change the display properties of regions (or XLDs) other than the topmost. What you can do is rebuild
the image stack in the graphics window manually by double-clicking iconic variables in the variable window and
changing the properties each time another layer is added. The stack is cleared whenever an image is added that
uses the full domain. To clear the stack (and the graphics window) manually, click , see figure 6.10 on page 74.
Figure 4.7: Determining threshold bounds interactively using the gray histogram.
Threshold
Range Selection and Code Generation
the gray histogram is initially empty. Double-click the Clip icon in the variable window to re-display the original
image and watch its gray histogram appear.
Make sure Range Selection and Code Generation is visible in the histogram window, see figure 4.7. Select
Threshold in the column Operation of the gray histogram window, and click the icon next to Threshold to
visualize the operation. Now, you can try different threshold bounds by altering the values in Min and Max or by
dragging the green and the red line in the histogram area. Any changes to these values are immediately visualized
in the active graphics window. The values 0 and 56 seem suitable for the lower and upper bounds, respectively.
4.9 Editing Lines 37
To edit a line in the operator window, double-click it in the program window. If you make changes to the parameters
and click OK or Replace, the original line in the program is updated. You can also edit the program directly in the
program window (see section 6.17.2 on page 128).
Double-click the second line of the program to adjust the threshold operation. Replace the value 30 with 56 and
click Replace. The program line is updated in the program window.
The last editing step was just a tiny modification of the program. Often, after editing many lines in your program
with perhaps many changes to the variables, you want to reset your program to its initial state and run it again to
see the changes.
Click Execute . Reset Program Execution to reset the program.
Now, you can select Execute . Run to run the complete program, or click Execute . Step Over repeatedly to
execute the program line by line.
The processing time of the operator or procedure call is indicated in the status bar at the bottom (unless disabled).
Additional Information
Programming
Note the following when executing a program via Execute . Run:
• The program line marked by the PC is the first line that is executed. All following program lines are going
to be performed until the end of the current program. After the execution is finished, the main procedure
becomes the current procedure.
• A breakpoint, stop instruction, or runtime error may interrupt the execution of your program; see below.
• If the HDevelop program waits for the user to draw something in the graphics window, a notification message
is printed in the status bar. The program halts until the user finishes the draw operation and confirms this
with a right click.
Further, the following characteristics apply during the execution of operator or procedure calls:
• You can initiate limited activities. For example, if you double-click variables in the variable window
they will be visualized. You can modify parameters for the graphics windows as described in the menu
Visualization. You can even modify the current procedure body.
• All user interaction except Stop is disabled during program execution if the latter was not started in the main
procedure. HDevelop may respond slowly to your actions while the program is running. This is caused by
the fact that HALCON reacts to user input only between calls to operators.
• A variable window update during runtime will only be performed if it has not been suppressed (see sec-
tion 6.16.14 on page 125). In any case, the values of all variables are shown in the variable window after the
termination of the execution.
• While the program is running, the menu items Run, Step Over, Step Into, and Step Out (and the corre-
sponding tool bar buttons) cannot be executed.
• Suggestions in the menu Suggestions are determined for the recently executed operator.
• After executing a program (regardless whether via Run or stepwise via Step Over), HDevelop is available
for further transactions. Any user input that has been made during execution is handled then.
A running HDevelop program stops if one of the following conditions is met:
• You click Execute . Stop (or the corresponding tool bar button).
• The last operator or procedure call in the current program (the main procedure body) is called. This is the
usual way a program terminates. The PC is positioned behind this operator.
38 Programming With HDevelop
• A breakpoint is reached (see section 6.17 on page 127). In this case, the last operator or procedure call that
will be executed is the one before the breakpoint. In case of a breakpoint on a variable, the program stops
after the program line that modified the corresponding variable (see also section 6.21.1 on page 162).
• A runtime error occurs. An input variable without a value or values outside a valid range are typical reasons.
In this case, the PC remains in the line of the erroneous operator or procedure call.
• A stop instruction is executed. The PC remains on the line containing the stop instruction. Note that stop
instructions inside locked procedures (see section 5.6 on page 47) are obeyed. However, the code of the
locked procedure will only be visible if the correct password is entered in the program window.
• The program waits via the operator wait_seconds. In this case, clicking Run, Step Over, or Step
Forward continues the program immediately.
• HDevelop is closed.
The procedure and procedure call in which the execution of a program was stopped automatically become the
current procedure and procedure call.
Perhaps now is a good time to save your program. Select File . Save and specify a target directory and a file
name for your program.
Inspecting the variable ConnectedRegions after the changed threshold operation yields a much better result.
Still, a contiguous area at the left edge of the image is returned ; see figure 4.8. To obtain only the regions that
coincide with the clips, we need to further reduce the found regions based on a common criterion. Analogous to
the gray histogram tool, which helps to select regions based on common gray values, HDevelop provides a feature
histogram tool, which helps to select regions based on common properties or features.
Click Visualization/Tools . Feature Histogram to open the tool. Make sure Feature Selection and
Code Generation is visible in the histogram window . The column Feature allows you to select the feature
that the region selection will be based on. The default feature is “area”, which is adequate in this case: The actual
clips are all the same size, thus the area of the regions is a common feature. In the feature histogram the horizontal
axis corresponds to the values of the selected feature. The vertical axis corresponds to the frequency of certain
feature values.
Similar to the gray histogram window, you can visualize the selected regions. The regions whose area falls
between the values Min and Max, which are represented by the green and red vertical lines, respectively. Click the
icon next to the selected feature (area) to enable the visualization.
Specify the parameters in the Input and Output section of the feature histogram window . Drag the green
and red line to see how this affects the selected regions. In the histogram we can see that in order to cover all
the clips, we can safely select regions whose area is between, say, 4100 and the maximum value in the histogram.
When you are satisfied with the selection, click the button Insert Code. The following line (with similar numeric
values) will be added to your program at the position of the IC:
Run the program, and inspect the output variable SelectedRegions. The regions corresponding to the clips
are now determined correctly. To obtain the orientation and the center of gravity of the clips, add the following
operator calls to the program:
Figure 4.8: Selecting regions with a similar area in the feature histogram.
Programming
Input and Output
Click to enable visualization
The operator orientation_region returns a tuple of values: For each region in SelectedRegions a corre-
sponding orientation value in Phi is returned. The operator area_center in the same way returns the area, row
and column of each input region as tuples. Again, run the program and inspect the calculated control variables. You
can inspect multiple control variables in one inspection window. This is especially useful if the control variables
all relate to each other as in this example. In the variable window select all control variables (hold down the Ctrl
key), and right-click Inspect (see figure 4.9).
Up until now, the visualization of iconic results relied on the fact that a graphics window is always opened by
default when HDevelop starts. If you need control over the size and position of graphics windows you can explicitly
open a distinct number of floating graphics windows.
Add the following lines before the first line of your program:
dev_close_window ()
dev_open_window (0, 0, 512, 512, 'black', WindowHandle)
The first line closes a floating graphics window potentially left over from a previous run to ensure a consistent
state at the start of the program. The second line opens a single graphics window that can be referenced by the
output variable WindowHandle. The window handle is a magic value that is also displayed in the title bar of the
corresponding graphics window.
Graphics windows that are opened via operator dev_open_window will be opened as floating windows, displaying
images in Full Stretch mode for maximum backwards compatibility.
1. Prepare the state of the graphics window to use. For more information, see section 6.8 on page 73.
2. In the program window, place the cursor where you want to insert the code.
3. Click Visualization . Insert Code....
4. In the dialog, specify the settings to add:
5. Click Insert.
Being an integrated development environment, HDevelop provides features found in other programming languages
as well: Variable assignment, expressions, and control flow. Variable assignment and control flow are implemented
in terms of specific HDevelop operators. These operators can be selected from the menu Operators . Control.
Expressions are implemented in terms of a specific HDevelop language which can be used in input control param-
eters of operator calls.
To iterate over the elements in Phi, we use a for loop which counts from zero (the index of the first element of
a tuple) to the number of elements minus one. The for loop is entered just like a common HALCON operator:
Enter for into the operator window and specify the parameters as in figure 4.10. Note that the closing endfor is
entered automatically if the corresponding check box is ticked. Also note that the IC is placed between the added
lines so that the body of the loop can be entered.
The notation |Phi| - 1 is part of the HDevelop language. This operation calculates the number of elements in
Phi minus one. When inserted in the program window, the operator for is displayed in a different format to make
it more readable.
4.15 Summary 41
Add the following instruction to the program. It is automatically indented in the program window to highlight the
nesting inside the for loop. The backslash at the end of the first line allows the program line to continue at the
next line in the program window for readability. You can either enter the instruction as displayed or in a single
long line without the backslash.
dev_disp_text (deg(Phi) + ' degrees', 'image', Row, Column, 'black', [], [])
Programming
4.15 Summary
This is basically the way to create programs in HDevelop. Select an operator, specify its parameters, try different
settings using the button Apply, add a new program line using Enter or OK, and edit it later by double-clicking
it in the program window. Use the interactive tools provided by HDevelop to assist you, for example, to find
appropriate values for the operators.
Chapter 5
HDevelop Procedures
HDevelop offers a mechanism for the creation and execution of procedures. Procedures are meant to increase the
readability and modularity of HDevelop programs by encapsulating functionality of multiple operator calls in one
or more procedure calls. It also makes it easier to reuse program code in other HDevelop programs by storing
repeatedly used functionality in external procedures.
An HDevelop procedure consists of an interface and a program body. Procedure interfaces resemble the interfaces
of HALCON operators. They contain parameter lists for iconic and control input and output parameters. The
procedure body contains a list of operator and procedure calls. Furthermore, HDevelop provides extensive support
to supplement procedures with structured documentation. The documentation is automatically integrated into the
online help system.
Every HDevelop program is made up of one or more procedures. It always contains the main procedure, which has
a special status inside the program, because it is always the top-most procedure in the calling hierarchy and cannot
be deleted from the program.
HDevelop offers all necessary mechanisms for creating, loading, deleting, copying, modifying, saving, and export-
Procedures
ing procedures. Once a procedure is created, it can basically be used like an operator: Calls to the procedure can be
added to any program body and be executed with the appropriate calling parameters. Generally, the concept of us-
ing procedures inside HDevelop is an extension to the concept of calling HALCON operators since procedure and
operator interfaces have the same parameter categories, and the same rules apply for passing calling parameters.
HDevelop supports different types of procedures. The type is specified when a procedure is created, and determines
the location of the procedure in the file system.
Each program file contains exactly one main procedure and zero to multiple local procedures.
Local procedures are stored inside the HDevelop program and cannot be called from other programs or exter-
nal procedures.
External procedures are stored as separate files, and can be shared between different HDevelop programs.
A modification immediately affects all HDevelop programs using it when reloaded. Each procedure file
contains a single external procedure. The file name determines the name of the procedure. Thus, if the file
name is changed, programs and other procedures using it will have to be adapted.
Libraries contain a collection of (typically related) procedures in a single file. They share the same properties
as a collection of external procedures in a single directory. The idea of creating a library is to keep related
procedures as a unit.
If the type of an existing procedure is changed, the procedure resolution might be affected (see section 5.5 on page
46).
44 HDevelop Procedures
HDevelop programs, procedures, and libraries are stored in files with different formats and extensions.
When a file is changed outside of HDevelop, the file tracking alerts the change and offers to reload the file. Click
Ignore to abort. To reload the changed files without using the file changes dialog, click Reload All within
the File tabcard. For more information, see section 6.16.12 on page 123.
Reload All
Ignore
.hdev This is the default file format for HDevelop programs. The “main” procedure and all local procedures. It
stores programs in XML format and is suitable for revision control software.
.dev (Legacy) This is the default file format for HDevelop programs in HDevelop versions up to and including
HALCON 9. It is not suitable for managing programs using revision control software. This format is
required if you want to be able to load HDevelop programs in older versions of HDevelop. This format is
now marked as legacy and its use is no longer recommended.
.hdvp This is the default file format for external procedure files. It stores external procedures in XML format and
is suitable for revision control software. Procedures with the extension .hdvp always override procedures
with the same name but the extension .dvp in the same directory.
.dvp (Legacy) This is the default file format for external procedure files in HDevelop versions up to and including
HALCON 9. It is not suitable for managing external procedures using revision control software. This format
is required if you want to be able to use external procedures in older versions of HDevelop. This format is
now marked as legacy and its use is no longer recommended.
The file name (without the extension) determines the name of the contained external procedure (see also section 5.1
on page 43).
5.2.4 Libraries 45
5.2.4 Libraries
.hdpl This is the file format for procedure libraries. It stores libraries in XML format and is suitable for revision
control software.
The scope of a procedure defines its visibility to other procedures. If a procedure is visible, it can be called and
thus executed. The scope can be either private or public.
In private scope, the procedure is only visible from procedures in the same directory (external procedure) or
the same file (program or library). The scope of local procedures is always private to the current program,
and external procedures can never see them.
HDevelop looks for procedures in a set of locations in the order specified in this section. Locations can be either
directories or library files.
1. Import locations
Import locations are specified in the program code using the import statement. Program lines following the
import statement may call procedures from that location.
For example, if you want to call the procedure config.hdvp from the directory
C:/Users/Public/procedures/common, here is how this can be achieved:
Procedures
...
import C:/Users/Public/procedures/common
config()
...
An import location is valid from its import call to the end of the procedure.
2. Session locations
HDevelop can be started from the command line with a set of locations that are searched for procedures in
the current session only.
Calling
adds the locations given in path name(s) to the list of searched procedure locations. Multiple locations may
be specified by using the system-dependent separator, “;” on Windows systems and “:” on Linux systems.
Subdirectories of the specified locations are searched recursively.
3. Static user-defined locations
Arbitrary locations can be added or removed in the preferences of HDevelop. Just like the standard pro-
cedure path below, each user-defined directory can be enabled or disabled independently. The user-defined
directories and their settings are persistent. They survive a restart of HDevelop. It is good practice to put
commonly used procedures into these locations.
Subdirectories of the specified locations are searched recursively.
46 HDevelop Procedures
1. Imported procedures
HDevelop looks for imported procedures (section 5.4). Note that for several import statements that contain
procedures of the same name, the last import call has priority.
2. Context of the calling procedure
If no imported procedure is found, the context of the calling procedure is significant for the procedure
resolution. The context depends on the type of the calling procedure:
• If the calling procedure is a local procedure, it looks for local procedures.
• If the calling procedure is a procedure file, it looks for procedure files in the same directory.
• If the calling procedure is a library procedure, it looks for procedures in the same library.
3. External procedures
HDevelop looks for external procedures (section 5.4) in the following order:
The tool tip of a procedure call reveals the location of the resolved procedure.
Resolution Example
As an example, we look at a program with three local procedures. Two procedure directories are defined (disre-
garding the standard procedure path).
+ C:/Users/Public/
|
+--+ example.hdev
| |
| +--- main
| +--- init
| +--- compute_results
|
5.6 Protected Procedures 47
+--+ procedures/project/
| |
| +--- init.hdvp public
| +--- local.hdvp private
| +--- setup.hdvp public
| |
| +--+ visualization.hdpl library
| |
| +--- init private
| +--- process private
| +--- setup public
|
+--+ procedures/common/
|
+--- config.hdvp public
Procedures can be protected by a password. Protected procedures may be executed by all users. However, the
interface, documentation, and program code can only be accessed if the correct password is supplied.
Procedures
Instead of protecting single procedures individually, all procedures of an HDevelop program or a library, respec-
tively, can also be protected as a whole.
Protected procedures alter between two states:
A protected procedure is locked if the password has not been entered in the current session. Locked procedures
cannot be modified, and the program code is not visible in the program window.
A protected procedure is unlocked after the correct password has been entered. Unlocked procedures may be
modified, and the program code is visible in the program window.
HDevelop procedures can be documented like operators. The documentation may include a detailed description of
the functionality of the procedure, example code, links to other procedures or operators, and concise documentation
of each parameter.
To manage a large collection of procedures, the procedures can be ordered in a hierarchical way. Procedures can
be ordered by chapters and sections just like operators.
The description of procedures can be formatted using Markdown syntax.
See section 6.17.8 on page 142 for details.
48 HDevelop Procedures
Certain HDevelop procedures are obsolete. These procedures are only provided for backward compatibility. They
are listed in the corresponding Legacy chapter and, just as for operators, a warning icon is displayed as a reminder
in the operator window as well as in the program window. If you move the mouser cursor over the latter one you
get a tool tip with corresponding warning message.
You can also add warning messages with corresponding icons yourself. For more information, see section 6.17.8.1
on page 142.
HDevelop supports just-in-time (JIT) compilation of procedures for optimized performance of HDevelop pro-
grams. The JIT compiler is enabled in the “experienced user” settings of the preferences dialog (see section 6.16.12
on page 123). Once enabled, JIT compilation works without any user intervention. Before executing a procedure
for the first time, HDevelop automatically compiles it to optimized bytecode. The additional compilation time is
only relevant when executing a procedure for the first time in the current session. A re-compilation is triggered by
changes to the corresponding procedure.
Messages of the JIT compiler are logged to the output console (see section 6.1.9 on page 63). This includes infor-
mative events to indicate that a procedure has been (re-)compiled as well as warning events to indicate problems
related to JIT compilation.
The speed-up that can be achieved by JIT compilation depends on the structure of the procedure code. If the
procedure body contains only consecutive operator calls, the performance gain will be negligible. However, if
the execution of a procedure includes a considerable amount of program line switching (for example, loops), the
performance gain can be significant.
When using F7 during program execution to step into a procedure, the procedure is executed uncompiled by the
HDevelop interpreter.
If an error occurs inside a compiled procedure, the whole procedure call is aborted. The program counter will not
indicate the actual program
line that caused the error. To debug the corresponding procedure you can step into the
procedure call with F7 (thereby interpreting its code) or disable JIT compilation altogether in the preferences.
Error message dialogs within try-catch blocks are always suppressed in JIT-compiled procedures regardless of
the corresponding setting in the “experienced user” preferences (see section 6.16.12 on page 123).
In HDevelop, the HALCON system parameter “use_window_thread” (see set_system) is activated by default to
ensure that on windows systems all top level HALCON graphics and text windows are opened in a special window
thread. This setting must not be deactivated. Otherwise, HDevelop may hang, for example, when displaying
objects in a HALCON graphics window during regular execution and the window was previously opened during
JIT-compiled execution.
Limitations
JIT compilation is not supported for procedures that use any of the following features:
5.9 Just-in-Time Compilation 49
If one of these features is found, the corresponding procedure is called uncompiled as before by the HDevelop
interpreter.
Note that when an exception is thrown in compiled procedures, the data slot “call_stack_depth” is always returned
as -1 (either when inspecting the exception in the variable window or using dev_get_exception_data).
Procedures
50 HDevelop Procedures
Graphical User Interface 51
Chapter 6
GUI Reference
Figure 6.1: HDevelop main window.
The HDevelop main window contains the following components: The window title shows the given name of the
current program. Unsaved changes in the current program are indicated with a trailing asterisk (*) in the window
title.
• Menu
The menu at the top provides access to the functionality of HDevelop. The menus and their entries are
described in section 6.1.
At the right end of the menu, there is a feedback button .
• Tool Bar
The tool bar icons provide convenient shortcuts for frequently used functions, see section 6.2 on page 64.
• Status Bar
The status bar displays context-sensitive information about a specific user action, or the runtime of operator
or procedure calls (unless time measurement has been deactivated in the preferences, see section 6.16.14 on
page 125).
For some very time-consuming operators, like trainf_ocr_class_mlp or train_class_svm, the status
bar also displays a progress bar for more information see section 6.3 on page 65.
52 Graphical User Interface
• Window Area
The main part of the window is reserved for the dockable windows and dialogs of HDevelop, like the Canvas
Window, the File Selection Dialog, the Graphics Window, and so on.
6.1 Menu
The menu of the main window provides access to the complete functionality of HDevelop. Every menu item opens
a drop-down
menu with optional submenus. You can either access the menu by clicking an item, or by pressing
the Alt key in combination with the underlined letter of the menu item. In the following sections, the menu entries
are described in the order in which they appear.
This menu provides functions to load existing programs and to save recently created or modified programs and
procedures. See section 5.2 on page 44 for the supported file types. Here, HDevelop programs can be exported to
C++, C, Visual Basic.NET, C#, or plain text and also be printed. The menu also provides access to the supplied
example programs and allows you to read arbitrary images.
The file type of programs and external procedures is persistent: If you load a program in the older .dev format
and save it again, it will not be converted to the newer .hdev format unless explicitly specified in the dialogs Save
Program As... or Save Procedure As.... Additionally, it is possible to convert HDevelop programs and
procedures between the old and the new format by calling hdevelop -convert from the command line.
For new programs, the default file format (.hdev) will be used. When you save a program for the first time, you
can also select the older file format in the corresponding dialog. If you want to use the older format all the time,
you can make it the default by modifying the preferences, see General Options (page 121).
GUI Reference
54 Graphical User Interface
This menu holds the functions to modify the current HDevelop procedure body as shown in the program window.
You can also access the preferences of HDevelop from this menu.
This menu item provides all necessary functions to execute an HDevelop program. In HDevelop, program execu-
tion is always continued at the top-most procedure call, which in most cases corresponds to the current procedure
call. The procedure body displayed in the program window belongs to the current procedure.
GUI Reference
Activate / Deacti-
mode. To deactivate/activate the breakpoints in many procedures Shift+F10
vate Breakpoint
at once, use the tab Breakpoints of the Quick Navigation Win-
dow (see section 6.18.3 on page 154).
Clear All Break- Both line-based breakpoints as well as breakpoints on variables
points are cleared.
Opens the Quick Navigation Window to manage breakpoints. Ctrl+Shift+O,F10 or
Manage Breakpoints
More info (page 154) Ctrl+Shift+O,Ctrl+Shift+F10
Reset Program The PC is set to the first executable line of the main procedure.
Execution All variables are reset. F2
Useful for debugging procedures without affecting the calling
Reset Procedure procedures. The PC is set back to the first executable line of
Execution the current procedure. All variables of the current procedure are Shift+F2
reset.
Abort Procedure The PC is set back to the first executable line of the current pro-
Execution cedure. All variables of the current procedure are reset. Shift+F8
Toggle the profiler section of the program window, which dis-
Activate / Deacti- or
Ctrl+Shift+F,F
plays runtime statistics, like the total and average processing
vate Profiler Ctrl+Shift+F,Ctrl+Shift+F
time of each program line. More info (page 148)
Select what is being displayed in the profiler section of the pro-
Profiler Display
gram window. More info (page 149)
56 Graphical User Interface
Via this menu, you can open or close graphics windows and clear their displays. Their output behavior during
runtime can also be specified here. Most functions are also available from the context menu of the graphics
window.
GUI Reference
specifies the update behavior when stepping through the pro-
gram.
Always: display iconic results
Update Window
Never: do not display iconic results
Clear And Display: clear graphics window before displaying
iconic results
Last Image And Display: keep displaying the latest image
and additional iconic results
As in Run Mode: same behavior as run mode setting
Select the precision of subpixel mouse positions. By default,
mouse positions are displayed as integers (precision 0), where
the upper left image pixel is displayed as 0, 0. Increasing the
precision results in mouse positions being reported as subpixel-
Position Precision
precise positions. Please note that when subpixel mouse posi-
tions are enabled, the position 0.0, 0.0 refers to the center of
the upper left pixel. The upper left edge of the image is displayed
as -0.5, -0.5.
To easier visualize the pixels when an image is zoomed, a Pixel
Pixel Grid
Grid can be set.
58 Graphical User Interface
Submenu Tools
This menu contains all functionality that is needed to create, modify, copy, or delete HDevelop procedures. To
save procedures, refer to the File menu (section 6.1.1 on page 52).
GUI Reference
serts only the procedures that are not locked. For changing the
status of a procedure, see section 5.6 on page 47.
Convert all local procedures into external procedures. Former
local procedures are stored as external procedures in a selectable
Make All External
directory of the list of external procedure directories. More info
(page 115)
Manage Proce- Opens the dialog Preferences . Procedures to configure pro-
dures cedure settings. More info (page 117)
Select a procedure for editing in the program window. This sub-
Edit Procedure menu lists all procedures in submenus, grouped by chapter and
section title. More info (page 142)
60 Graphical User Interface
This menu contains all HALCON and HDevelop operators including the HDevelop control constructs as well as
all internal and external HALCON procedures.
For detailed information about all operators and procedures, we recommend reading the corresponding sections
of
the reference manual. To get there quickly, select an operator or procedure from the menu, and then press F1 .
This menu shows another possibility how to select HALCON operators via suggestions. Assuming that you have
already selected an operator in a previous step, depending on this operator, the following suggestions are offered.
Suggestions are separated into groups as described below.
GUI Reference
62 Graphical User Interface
This menu assembles assistants for specific machine vision tasks. The general concept of the assistants is described
in the chapter Assistants (page 179).
The following assistants are available:
This menu offers support to manage the sub-windows of the main window, such as the program, operator, variable,
graphics window(s), and possible other dialogs.
At the bottom of the menu, all open windows are listed. Click an entry to bring the corresponding window to the
front.
GUI Reference
Bring Others To Raises all floating windows to the top. Works only if your win- or
Ctrl+Shift+W,B
Front dows are minimizable. Ctrl+Shift+W,Ctrl+Shift+B
or
Ctrl+Shift+W,F
Full Screen Toggle full screen mode.
Ctrl+Shift+W,Ctrl+Shift+F
Restore Default Lay- The canvas and the docked graphics window are docked top left.
out Use
64 Graphical User Interface
This menu offers a version check and access to the HALCON documentation.
Status icon
Shows the current run status of the program. : Program stopped, : Program executing.
Messages and runtime information
For example, if you select an operator from the menu, the corresponding short description is displayed here.
The runtime information depends on the run mode:
• When single-stepping through the program, the runtime of the last operator or procedure call is dis-
played. If the option Show memory usage is activated in the preferences (see section 6.16.12 on
page 123), the operator’s required temporary HALCON memory will be displayed additionally. The
displayed memory does not include objects created by the operator, like images or tuples.
• In continuous run mode, a runtime summary of the executed program lines is displayed when the
program stops.
Right-clicking in the message area opens the context menu. It provides the following entries:
• Show Processing Time: Toggles whether execution messages are displayed in the status bar.
• Copy History to Clipboard: A history of the latest execution messages is displayed as a tool tip
when placing the mouse pointer over the message area of the status bar. The history can be copied to
the clipboard by selecting the entry Copy History to Clipboard in the context menu of the status
bar.
• Open Output Console: The output console displays the message history in a separate window. See
section 6.1.9 on page 63.
GUI Reference
Information about the image in the active graphics window
The display format is [index] variable name (#=number of objects: height x width x channels x type)
Gray value of the image in the active graphics window at the mouse cursor position
For multi-channel images, the gray values of all channels are displayed separated by commas.
MVTec logo
Click the logo to open our website http://www.halcon.com.
66 Graphical User Interface
• Auto-fit By Zooming In
The view is zoomed in automatically, so that all graphics windows on the canvas are visible at a glance with
optimal use of the available space. This means, as soon as you move a graphics window towards the center
of the current view, the view adapts automatically by increasing the zoom factor to maximum 100%.
• Auto-adapt Viewport
The canvas view is focused automatically on the selected graphics window. Thus, the view follows the
selected graphics window. To scroll the canvas view manually, disable this option.
• Show Grid
The grid helps you arranging and resizing your graphics windows. The cross in the grid visualizes the
reference point (0, 0) for the graphics window position.
• Show Graphics Window Toolbar
You can hide the toolbar, for example, to maximize the image display. Most graphics windows features are
available via context menu, too. For example, the feature Float window to remove the graphics window
from the canvas again.
GUI Reference
• Show Graphics Window Frame
The frame is necessary to make changes to the size or position of the graphics window. Disable this option,
if you are happy with its current position and size or if you want to focus on the image.
68 Graphical User Interface
6.4.2 Zooming
1. Click inside the canvas window (either in the background area or inside a graphics window on the canvas).
→ The canvas window gets the focus, indicated by its frame turning from gray to orange.
2. Press Ctrl+Shift++ to zoom in, or press Ctrl+Shift+- to zoom out. Alternatively, hold Ctrl or Ctrl+Shift and
i scroll up to zoom in, or scroll down to zoom out. You can change the zoom direction of the mouse
wheel via the general option Action when spinning the mouse wheel down (page 121). Note that this
setting also applies to the graphics window.
→ The current zoom factor is displayed in the bottom right corner of the canvas window.
Click the zoom factor to reset the view to 100%.
6.4.3 Moving
6.4.4 Resizing
1. Enable the canvas option (page 66) Show Graphics Window Frame.
2. Drag (page 12) the bottom right corner of the graphics window.
! Resizing is not possible during draw operations, like draw_rectangle1.
6.5 Browse Example Programs Dialog 69
Category name
Find text field
Browse the categories: Click a topic to select it and display its example programs. You can select multiple
topics at once by holding the Ctrl key while clicking the categories.
Filter the example programs: To reduce the amount of listed example programs, enter a word or substring into the
Find text field. Only example programs matching this substring in the file name or short description will be
displayed.
Assume you are looking for a measuring example from the semiconductor industry:
GUI Reference
1. Double-click Industry.
2. Click the subtopic Semiconductors. The examples belonging to the semiconductor industry are listed on
the right.
3. Enter the word measure into the Find text field.
Note how the list is updated as you type. Now, you have a short list of example programs to select from.
You may need to resize the example browser to fully read the short descriptions of the listed programs.
4. Select measure_ic_leads.hdev by clicking it.
5. Click Open. The selected example program is then loaded. Alternatively, you can load an example program
by double-clicking it. The example browser is closed unless Keep dialog open is selected.
Export Range
The export range specifies which parts of the current program are to be exported.
• Program: The entire program is exported the main procedure and all local procedures. All used exter-
nal procedures are exported depending on the setting of the external procedure options (see below).
• Current Procedure: The current procedure and all used local procedures are exported. All used ex-
ternal procedures are exported depending on the setting of the external procedure options (see below).
• External Procedures: All external procedures are exported depending on the setting of the external
procedure options (see below).
• Current Library: The current library is exported (all procedures that are part of the current library).
All used external procedures (that are not part of the current library) are exported depending on the
setting of the external procedure options (see below).
The short description and chapter information of procedures are exported as comments. Arbitrary code can
be embedded with special comment lines (see section 10.2.5 on page 314).
Procedure Options
Defines the export behavior for procedures.
• Export local procedures body: If checked, both the declaration and the body of local procedures
are exported. Otherwise, only the declaration is exported.
• Export external procedures body: If checked, both the declaration and the body of external
procedures is exported. Otherwise, only the declaration is exported.
• Export all referenced external procedures: Determines, if all referenced external proce-
dures are also exported with the current program.
• External Procedures Destination: By default, external procedures are exported to a single file.
Select Separate Files to export external procedures to separate files. The file name corresponds to
the procedure name while the file extension is derived from the export format. Select Separate or
Library Files to export external procedures to separate files but keep the files of libraries together
in a single file.
• Export only if procedure is newer than export file: Export only those procedures to
separate files whose time stamp is newer than the time stamp of the destination file.
Window Export
Specifies the export behavior of HALCON windows:
• Use HALCON Windows: Export as a stand-alone project.
• Use Export Template (HALCON/.NET only): Export as a project using the supplied project tem-
plate.
Encoding
Specifies the encoding of exported programs.
• Native: Export in the encoding defined by the operating system.
• UTF-8: Force export in UTF-8 encoding (Unicode).
For C/C++, the encoding must match the interface encoding used by the application. For C#, exporting to
UTF-8 should always work fine.
Namespace
Can be used to avoid name conflicts between exported code and existing names in the client application, or
GUI Reference
to avoid name conflicts between multiple exported libraries in the same client application.
• If this field is empty no namespace is used.
• If this field is filled all code, generated during this one export operation, will be wrapped using exactly
this one name.
• Template projects expect exported code without namespace, and a template export with namespace
only works with adapted projects.
• When performing multiple exports with different namespaces the exports for each namespace must
be self-contained including all dependencies. Exported procedure calls will not "see" exported pro-
cedures from the other namespace. The solution here would be to export the procedure call, with all
dependencies even if it can lead to code duplication.
• The separate exports cannot share global variables. To avoid this: Do not use global variables, or use
multiple "global def" declarations, and do not rely on sharing of global variables between libraries.
This icon appears in some cases, when an export may not be working. Hovering over the icon provides
more information.
The Find/Replace dialog provides comprehensive facilities for searching the program code. You can perform a
full text search or search for variable names as well as operator (or procedure) calls. In addition, you can replace
variable names and substitute operator or procedure calls. The dialog is displayed in figure 6.7.
You can even select multiple lines from the search result by holding the Ctrl key. The following actions may
be performed for all selected lines (from the context menu of the search result): Cut (page 54), Copy (page 54),
Delete (page 54), Activate (page 54), Deactivate (page 54), and Bookmark Selection. Use Bookmark
Selection to add bookmarks to the selected program lines. This will also add the selected program lines to
the tab Bookmarks. This way you can “remember” a search result for later use; just open the Quick Navigation
window again (Menu Window . Open Quick Navigation) and go to the tab card Bookmarks.
The “find all” operation is recommended before doing a global replace to preview which program lines will be
affected.
GUI Reference
Replacing Operator Calls
You can replace one operator or procedure call with another. Because different operators very likely have different
parameters, the source parameters have to be mapped to the target parameters beforehand. See figure 6.9 for an
example.
Click Operators to specify the search context. Enter the source operator or procedure name and the target operator
or procedure name. When both names are specified, the parameters of the target operator/procedure are listed at
the bottom of the dialog. For every target parameter you have to select or enter a corresponding source parameter.
Please note that the option Whole Words is always enabled in this mode because only exact matches are valid
when replacing operator calls.
Full view
The origin of the image coordinate system is the upper left image corner with the coordinates (0,0). The x values
(column) increase from left to right, the y values (row) increase from top to bottom. When the mouse cursor is
placed inside a graphics window, the coordinates (row, column) and the gray value (or in this case: the RGB values)
at that position are displayed in the status bar (see section 6.3 on page 65). Sometimes,
it is desirable to display
this information close to the mouse cursor. This can be achieved by holding down the Ctrl key.
! Please note this does not work when the zoom in and out tool is selected since pressing Ctrl inverts the
corresponding zoom action.
Figure 6.10 shows the coordinate/color value display after zooming in with the mouse wheel.
Normally, the visible part in the graphics window is based on the most recently displayed image, which is au-
tomatically zoomed so that every pixel of the image is visible. The coordinate system can be changed interac-
tively using the menu Visualization . Set Parameters . Zoom (see Zoom settings (page 78)) or the operator
dev_set_part. Every time an image with another size is displayed, the coordinate system is adapted automati-
cally.
Every HDevelop graphics window has its own visualization parameters. Thus, modifying the parameters (see
Visualization Parameters (page 76)) applies to the currently active graphics window only. The parameter settings
6.8.1 Tool Bar 75
Additional Information
When opening a graphics window via Visualization . Open Graphics Window... and the window height
and width are set to -1, the window size is set by HDevelop. It is taken from the persistent preferences of HDevelop
(usually the size of the last graphics window in the previous HDevelop session).
GUI Reference
Set image size. The value can be selected from the menu attached to the small arrow. Additionally, the resizing
behavior can be selected.
The following values for image scaling are available:
• Fit
Scales the image to completely fill the graphics window.
• 400 %, 200 %, ...
A list of fixed percentages scales the image with respect to its natural size.
• Double
Double the current image size.
• Half
Half the current image size.
• Aspect Ratio 1:1
Scales down the image size, so that its aspect ratio is maintained.
You can also change the size of the graphics window, for example, by “gripping” the window border with
the mouse. Then you can resize the window by dragging its border. After this size modification the window
content is redisplayed, depending on the chosen resize mode.
The following modes to control the resizing behavior of the image are available:
76 Graphical User Interface
• Full Stretch
The visible image part remains constant. The zoom levels are adapted to fully fit the previous view
into the new window size. The aspect ratio of that image part might be skewed.
This is the classic mode. Graphics windows that are opened via operator dev_open_window will be
opened as floating windows using the “Full Stretch” mode for maximum backwards compatibility.
• Keep Aspect Ratio
The zoom level is changed to preserve the previous aspect ratio. The visible image part can change in
one axis as a result of the resize operation.
This is the default mode. You can change the default mode under Edit . Preferences . General
Options.
• No Stretch
The zoom level remains fixed. Thus, resizing the window will show a correspondingly larger or smaller
portion of the image.
Set window size. Clicking this icon sets the window size to the shown value. The value can be selected from the
menu attached to the small arrow. Double and Half change the size to half and double its current window
size, respectively. Aspect Ratio 1:1 scales down the current window size, so that the aspect ratio of the
displayed image is maintained.
Draw ROIs and XLDs interactively; see section 6.19 on page 155.
If you want to specify display parameters for a window, you can select the menu item Visualization in the
menu. Here you can set the appropriate parameters by clicking the desired item (see section 6.1.4 on page 56).
The parameters you have set this way are used for the active window. The effects of the new parameters will be
applied directly to the last object of the window history and alter its parameters only.
Click Set Parameters (via the context menu or the menu Visualization) to open the Visualization
Parameters. This dialog allows convenient access to the visualization settings of the active graphics window.
Most of the settings are also available as individual menu entries in the menu Visualization but some more
advanced settings are only provided in this dialog. An interactive preview is provided, which visualizes the current
settings.
Use settings for new windows Make the current settings also the default settings for new graphics windows.
6.8.2 Visualization Parameters 77
The pen settings specify the drawing mode, filled or outlined, the shape, the line width, and the colors of regions
and XLDs.
For regions the draw mode can be set to filled (item fill) or outlined (item margin).
The parameter Shape, default is original specifies the presentation shape for regions. You can display not only
the region’s original shape but also its enclosing rectangle or its enclosing circle, etc.
The setting of Line Width applies to XLDs and regions in draw mode margin.
The color of the regions/XLDs can be set to one of the named colors. If you select one of the color sets, each region
or XLD will be displayed in an alternating color from a set of 3, 6, or 12 colors. This visualizes the connectivity
of different regions in the graphics window.
Click the “add” button to define additional colors. Colors are specified by hue, saturation, value, or by giving the
red, green and blue values (see figure 6.13 on page 79). User-defined colors are added to the list of named colors.
They can be removed by selecting them and clicking the “remove” button.
You can also define your own custom color sets: Add selected colors to the list of custom colors by clicking the
“right” button, or remove them by clicking the “left” button. The check box next to Custom colors: specifies if
the custom color set is used.
User-defined colors and custom color sets are discarded when you quit HDevelop. If you want them to be perma-
nent, you will have to define them in the preferences dialog (see section 6.16.13 on page 125).
The pen settings are also available from the corresponding menu entries in the menu Visualization. A descrip-
tion of the functionality is provided there. The preview shows the current settings, which is helpful if the active
graphics window does not contain any regions or XLDs.
• “Draw”, see also section 6.4 on page 56
• “Colored”, see also section 6.4 on page 56
• “Color”, see also section 6.4 on page 56
• “Shape”, see also section 6.1.4 on page 58
• “Line Width”, see also section 6.1.4 on page 58
Using LUT you are able to load different look-up tables for visualization. With the help of a false color presentation
GUI Reference
you often get a better impression of the gray values of an image. In the case of a true color display, the image has
to be redisplayed due to the missing support of a look-up table in the graphics hardware. For color images only the
gray look-up tables can be used, which change each channel (separately) with the same table.
See the description of the menu entry “Lut” in section 6.1.4 on page 58.
Here, you can select between two graphical image presentations. In the default mode the image is displayed
unmodified. In the 3d_plot mode, the gray values of the image are taken as height information: The greater the
gray value, the higher the resulting image point. See figure 6.16 on page 81 for an illustration of the different
modes. Further information can be found at the description of the operators dev_set_paint and set_paint.
This tab card lets you set a Pixel Grid to easier visualize the pixels, when zooming in. You cannot configure the
Pixel Grid within HDevelop using the operator set_window_param.
Predefined colors
Color sets
Add/remove user-defined color
6.8.3 Context Menu 79
GUI Reference
Figure 6.14: Custom color set.
Below the coordinates of the rectangle, you can specify its center.
The buttons Zoom Out and Zoom In activate a zooming with factor 0.5 or 2, respectively.
To get the image’s full view back on your graphics window, click the button Reset.
The button Aspect adjusts the parameters so that the aspect ratio of the image is maintained.
The context menu can be enabled/disabled from the preferences (page 125).
The entries of the context menu are a subset of the menu Visualization (page 56). Graphics windows at-
tached to the canvas contain the additional context menu entry Canvas Options. For more information, see i
section 6.4.1 on page 66.
80 Graphical User Interface
Clicking activates an interactive 3D plot mode. It displays meaningful information for height field images, such
as images that encode height information as gray values. The greater the gray value, the higher the corresponding
image point. Figure 6.19 shows a height field image of a solder ball and the corresponding 3D plot .
The 3D plot mode uses OpenGL and benefits from hardware acceleration.
Using the mouse you can alter the view of the 3D image (select mode must be active for this to work, click in
the tool bar):
There are four different rendering methods (texture, shaded, hidden_lines, and contour_lines) which can be selected
from the drop-down menu in the tool bar. See set_paint for detailed information about the different methods.
The display quality can be fine-tuned in the tool bar of the graphics window or in the visualization parameters
of the graphics window (right-click into the graphics window, select Set Parameters, and open the tab card
Paint).
Mode sets the rendering mode just like the drop-down menu in the graphics window.
Plot Quality allows setting the rendering quality in four steps. On systems without proper display hardware
acceleration a lower quality should be selected to speed up the display.
Step sets the level of detail. In general, the lower the step value, the higher the level of detail. However, if the
rendering mode is set to contour_lines, increasing the step value increases the level of detail.
Display Axes If this is enabled, the axes of the 3D coordinate system are displayed in the 3D view.
Display Grid If this is enabled, the “floor” of the 3D plot is painted as a grid.
See also section 6.8.2.3 on page 78 for the other paint modes that can be selected in this window.
If there is at least one docked graphics window, all floating graphics windows are closed with F2 . This shortcut
also works in the main window.
GUI Reference
More keyboard functions of the graphics window are listed in appendix D.2 on page 327.
The help window provides access to HALCON’s online documentation. The window is split into two areas: On
the
left, navigational
panels are available as tab cards. They are described below. Please note that you can press the
Up and Down keys in the navigational panels to select the previous and the next entry, respectively. This allows
you to examine, for example, search results quickly. The main area displays the actual hypertext content. The size
of the two parts of the help window can be adjusted by dragging the dividing line.
Parts of the documentation are available in PDF format. The help window does not display these files itself but
launches the default viewer when a PDF link is being followed. If a link to a PDF file appears in the navigation
(either as the result of a full-text search or from the selection of index keywords), it will be marked by a PDF icon.
A single click will select the entry and display a link to the file in the contents area. A double-click will open
the corresponding file in the external viewer. This way, you can quickly browse the search or index results in the
navigation without accidentally running the PDF viewer.
Contents
This tab card presents the chapters and sections of the online documentation as a hierarchical tree. Click a node of
the tree to display the associated document. See figure 6.21 on page 84 for an example.
82 Graphical User Interface
3D plot mode
Operators
This tab card lists all operators in alphabetical order. Click an operator name to display the corresponding page
from the Reference Manual. Enter any name into the text field Find to show only operators matching that name.
Search
Enter a search query into the text field, and click Find to start a full-text search. The check boxes below the text
field indicate the search scope. For example, to make sure only HTML matches are displayed, deselect the check
box PDF. The search result is displayed below the query. The rank (in percent) indicates how well each found
document matches the query.
The query can consist of one or multiple words. HDevelop will find all documents that contain any of the specified
words.
To search for a phrase, enclose it in double quotes:
"radiometric calibration"
Precede all mandatory words with a plus sign (+). For example, to find all documents that contain filter and gauss,
enter:
6.9 Help Window 83
Visualization parameters
Tool bar
Mode
Quality
Step
Axes
Grid
GUI Reference
+filter +gauss
You can exclude specific words from your search result. To find all documents that say anything about filters except
Gaussian filters, enter:
filter -gauss
Index
This tab card provides access to HALCON operators and relevant sections of the documentation through index
entries. The list of index entries can be filtered by entering a string into the text field Find. If you enter multiple
words, only index entries matching all the words are displayed.
When you select index entries from the list, the related operator names and links to the corresponding
parts
of the
documentation are displayed below the index entries. The subtopics can be navigated using Ctrl+Up and Ctrl+Down
to select the previous and the next match, respectively.
Bookmarks
This tab card lists all user-defined bookmarks. You can add the currently displayed document to the list by clicking
the button Add. To remove a bookmark from the list, select it and click the button Delete.
84 Graphical User Interface
GUI Reference
86 Graphical User Interface
Selecting this entry opens a tool for inspecting and modifying training files. With this tool, you can for example,
eliminate errors made in the teaching process, for example, if a sample has been assigned to a wrong symbol
(please follow the links for a definition of the terms symbol (page 232) and sample (page 232)).
This tool is a useful addition to the OCR Assistant (page 231) but can also be used independently for any OCR
application. It can be opened either via the corresponding toolbar button, from within the OCR assistant (page
231) or via the drop-down menu Visualization . Tools . OCR Training File Browser.
Note that if the OCR Training File Browser is used in combination with the OCR Assistant (page 231), the samples
are displayed inverted if Light-On-Dark has been chosen as Symbol Appearance within the assistant. This has
an effect on the zoomed sample visualization in the lower left window as well as the thumbnail view the right
window.
The following sections will introduce you to the OCR Training File Browser by
• providing an overview over the different windows within the Training File Browser,
• explaining typical steps of using the tool in an application,
• giving an overview over all possible actions (page 88) within the training file browser.
The OCR Training File Browser is composed of three windows (see figure 6.22). The upper left window lists
training files and symbols and is described in the section ’Training File Window’. Underneath this window, the
’Zoomed Sample Window’ allows you to view the magnified image of a selected sample. The ’Sample Inspection
Window’ on the right lists details for selected samples to inspect the training results.
This window of the OCR Training File Browser enables you to view and edit training files (page 88) and symbols
(page 88).
This window displays the symbol that is selected in the Sample Inspection Window. If a symbol has been selected
in the Training File Window, the first sample is displayed by default.
The Sample Inspection Window enables you to view and edit samples (page 89). It displays a table containing the
following information about each sample that belongs to a symbol or a training file that is selected in the Training
File Window.
The columns show the following:
• Image: iconic sample
• Symbol Name: class to which the sample was assigned
• Width: width of the iconic sample (in pixels)
• Height: height of the iconic sample (in pixels)
• Generated: whether the sample is a generated variation (page 90) or not
• Read Symbol: result of reading the sample with the selected OCR classifier
• Confidence: a measure for the reliability of the read symbol (0: no confidence, 1: highest confidence)
• Correct: whether a sample has been read correctly. true means that the values of Symbol and Read
Symbol are identical and false means that those values differ. If the sample has not been classified yet, the
value stays true.
6.10.2 Steps for Working With the OCR Training File Browser
This section quickly guides you through a possible application workflow with the OCR Training File Browser.
GUI Reference
More detailed information on the functionality of the OCR Training File Browser can be read in the sections
corresponding to each of the windows within the OCR Training File Browser.
1. Open the Training File Browser either via the corresponding toolbar button or from the OCR Classifier
tab within the OCR Assistant (page 231).
2. Load an existing training file or create a new training file via the corresponding toolbar buttons or the
corresponding entries in the drop-down menu File. If several training files should be used or inspected, all
of them can be loaded into the OCR Training File Browser.
3. If required, it is possible to add samples to the training file with the training file functionality of the OCR
Assistant (page 231) if required.
4. Choose a classifier via the drop-down menu Classifier . Load Classifier to classify your samples.
If you have an own or previously trained classifier, you can load it. Otherwise one of the trained OCR
classifiers provided by HALCON can be selected.
5. Inspect and/or edit the contents of training files. This can include adding new samples, deleting samples,
combining existing samples for a new training file or adding sample variations (page 90) as well as checking
for classification problems. Use the Training File Window to edit the training file and the Sample Inspection
Window to inspect and edit samples.
6. Save changes, for example, via File . Save Training File.
88 Graphical User Interface
7. Use the training file within an OCR application or continue to prepare an application with the OCR assistant
(page 231).
Note that all actions can be undone or redone by selecting Undo or Redo in the menu Edit to reverse or repeat
changes, respectively.
The Training File Window (page 87) provides the following options for editing training files:
• Open/Close training files: When opening the OCR Training File Browser via the OCR Assistant (page 231),
the current training file is automatically loaded. Otherwise, training files can be loaded or closed via the
corresponding entries in the menu item File or via the corresponding toolbar buttons.
• Create training files: Training files can be created via the entry New Training File in the menu File or
via the corresponding toolbar button.
6.10.3.2 Symbols
The Training File Window (page 87) provides the following options for editing symbols:
• Add a symbol
To add a new symbol to a training file, either
• press the toolbar button Add Symbol Name or
• select the menu entry Edit . Add Symbol Name.
• Delete a symbol
Select the symbol you want to delete and either
• press Del on your keyboard or
• select the menu entry Edit . Delete.
• Rename a symbol
To assign the samples to another symbol, just edit the symbol name. This action is equal to moving the
symbol (and therefore all samples that are assigned to this symbol) via drag and drop to another symbol.
Note that it is also possible to assign single samples to a new symbol. To learn more about assigning samples
to a new symbol class, read the information about modifying samples in the paragraph ’Improve your training
file’ in the section Sample Inspection Window (page 87).
• Copy a symbol
A symbol and all samples that are assigned to this symbol can be copied to another training file. To copy and
paste a symbol, either
• drag and drop the symbol while pressing Shift or
• press Ctrl+C followed by Ctrl+V , or
• select the menu entry Edit . Copy followed by Edit . Paste.
• Move a symbol
A symbol can also be moved via drag and drop
• to another symbol in the same training file,
• to another symbol in a different training file, or
• it can also be added to a different training file by moving it there.
Moving a symbol (and therefore all samples that are assigned to this symbol) to another symbol is equal to
assigning a new symbol name explicitely.
6.10.3 Actions Within the Training File Browser 89
6.10.3.3 Samples
The Sample Inspection Window (page 87) provides the following options for viewing and editing samples:
GUI Reference
Select one or more samples in the Sample Inspection Window (page 87). They can then be moved in the
Training File Window (page 87) via drag and drop to
• another symbol in the same training file or to
• another symbol in a different training file
• Delete a sample
Select one or more samples in the Sample Inspection Window (page 87). They can then be deleted
• by pressing Del on your keyboard or
• via the menu entry Edit . Delete.
Note that you can also add variations for samples via the Generate Sample Variations dialog that can be
opened via the corresponding toolbar button.
• Use the checkbox Display only incorrectly classified samples to view only samples that were
classified incorrectly.
• Sort samples by Confidence, as this may provide a hint to classification problems.
• Edit Symbol Name to assign a sample to the correct symbol class if necessary. (This action is equal to
moving a sample via drag and drop to another symbol in the Training File Window (page 87).)
The best classification results are achieved if the OCR font is trained using real data from the target application.
This can however be time-consuming. To speed up the creation of a large number of different samples per symbol,
it is possible to vary existing samples.
To add variations for certain samples, select either
• one sample in the Sample Inspection Window (page 87) to generate variations for this sample or
• a certain symbol in the Training File Window (page 87) to generate variations for all samples that are
assigned to that symbol, or
• the whole training file can be selected in the tree within the Training File Window (page 87) to generate
variations for all samples within this training file.
Then open the Generate Sample Variations dialog either via the corresponding toolbar button or via the Edit
menu.
This dialog allows you to select several types of variations, depending on what is required for the application. In
general, we recommend using as many samples as possible. So, if you are in doubt whether a variation type applies
or not, you should select it.
The following types of variations can be selected:
The OK button starts the generation of the new samples. Note that the generation might take some time if lots of
variations are selected for a large number of samples. A progress bar shows how far the generation has proceeded.
The generated samples can subsequently be viewed and edited in the Sample Inspection Window (page 87) where
they are marked with true in the Generated column.
Note that once the training file is saved, the generated samples cannot be recognized as generated ones any more
and therefore the Generated column will show false even for a sample that has been generated.
This window is used to edit and display an operator or procedure call with all its parameters. Here you will obtain
information about the number of the parameters of the operator or procedure, the parameter types, and parameter
GUI Reference
values. You can modify the parameter values according to your image processing tasks. For this you can use the
values proposed by HDevelop or specify your own values.
The operator window consists of the following four parts:
• At the top you find the operator name field, with which you can select operators or procedures.
• The large area below the operator name field is called the parameter display; it is used to edit the parameters
of an operator or procedure.
• The section labeled Advanced Parallelization Options allows you to call operators or procedures as
a subthread (see section 8.11.1 on page 283).
• The row of buttons at the bottom allows you to control the parameter display.
92 Graphical User Interface
The
operator name field allows you to select operators or procedures by entering (part of) their name. After pressing
Return or pressing the button of the combo box, the system is looking for all operators or procedures that contain
the entered name. The order of the listed result is as follows: Operators and procedures whose names begin with
the given substring are listed first, followed by all operators and procedures that contain the substring elsewhere.
Both parts of the list are sorted in alphabetical order.
If there is an unambiguous search result, the parameters are displayed immediately in the operator window. If there
are several matching results, a combo box opens and displays all operators or procedures containing the specified
substring. By clicking the left mouse button you select one operator and the combo box disappears. Now, the
operator’s parameters are shown in the operator window.
The short description of the selected operator is displayed in the status bar. The operator name is displayed in the
window title of the operator window.
The parameter display is the main part of the operator window. If you have selected an operator or procedure call,
HDevelop displays its interface, with the name, value, and semantic type of each parameter.
You can switch between a one-column (default) or two-column layout of the parameter display. In the two-
column layout, the iconic parameters are shown on the left and the control parameters on the right.
In the following description, “column” refers to the order of the entries, which is independent of the layout.
• In the first column of the parameter display the parameter types are indicated by icons. Note that icons are
not repeated if a parameter is of the same type as its predecessor.
• In the second column of the operator window you find the parameter names.
• The third column consists of the text fields, which contain variable names in case of iconic and control
output parameters and expressions in case of control input parameters. If you want to change the suggestions
offered by the system (variable names or default values), you can do so either manually or by clicking the
arrow button connected with the respective text field. This opens a list containing a selection of already
defined variables and other reasonable values from the operator knowledge base. By clicking the appropriate
item, you set the text field and the list disappears.
For the operators open_framegrabber, set_framegrabber_param, and get_framegrabber_param, the
value list of certain parameters is dynamic: It depends on the selected image acquisition interface. An
even more reasonable parameter suggestion is given if the corresponding handle is opened. If this dynamic
behavior is undesired, it can be disabled in the preferences; see section 6.16.12 on page 123.
This column can also contain action buttons for special semantic types, for example, a button to browse the
file system for the parameters that expect a file name.
• The fourth column indicates the parameter’s default semantic type and, optionally, its data type in parenthe-
ses.
Please refer to the following rules on how parameters obtain their values and how you can specify them:
Iconic input parameters Possible inputs for these parameters are iconic variables of the corresponding type. If
there is no need to execute the operator or procedure call immediately, you can specify new variable names,
meaning, names, that do not already exist in the variable window, but will be instantiated later by adding
further operators or procedure calls to the program body. In any case, you have to specify iconic parameters
exclusively with variable names. It is not possible to use expressions.
Iconic output parameters These parameters contain default variables, which have the same names as the param-
eters themselves. If a variable with the same name as the output parameter is already instantiated, a number
is added to the name to make it unique. Because the parameter names characterize the computed result very
well, you can adopt these default names in many cases. Besides this, you are free to choose arbitrary names
either by yourself or by opening the list (see above). If you use a variable that already has a value, this value
is overwritten with the new results. It is possible to specify a variable both in an input and output position.
6.11.2 Parameter Display 93
Control input parameters These parameters normally possess a default value. As an alternative, you can use the
text field’s button to open a combo box and to select a suggested value. In addition, this combo box contains
a list of variables that contain values of the required type. A restriction of proposed variables is especially
used for parameters that contain data like file, image acquisition, or OCR handles.
Input control parameters may contain constants, variables, and expressions. Common types are integer
numbers (integer), floating-point numbers (real), boolean values (true and false), character strings
(string), and handles (handle).
You can also specify multiple values of these types at once by using tuples. This is an array of values,
separated by commas and enclosed in square brackets. You can build up expressions with these values. You
GUI Reference
can use expressions in HDevelop similar to the use of expressions in C or in Pascal. You will find a detailed
description in section 8.5 on page 253.
Control output parameters: These parameters are handled in the same way as iconic output parameters. Their
defaults are named as their parameter names. Other possibilities to obtain a control output variable name
are either using the combo box or specifying variable names manually. You cannot use any expressions for
these parameters.
After discussing what can be input for different parameters, it is explained how this is done. Nevertheless, you
have to keep in mind that you need to modify a parameter only if it contains no values or if you are not satisfied
with the suggested default values.
Text input: Give the input focus to a parameter field by clicking into it. Now, you can input numbers, strings,
expressions, or variables. There are some editing functions to help you doing input: Backspace deletes the
character to the left and Delete deletes the one to the right. You can also select a sequence of characters in
the text field using the mouse or holding Shift and using the cursor keys. If there is a succeeding input, the
marked region is going to be deleted first and afterwards the characters are going to be written in the text
field. See appendix D on page 327 for a summary of the keyboard mappings.
Combo box selection: Using this input method, you can obtain rapid settings of variables and constants. To do
so, you have to click the button on the text field’s right side. A combo box is opened, in which you can select
an item. Thus, you are able to choose a certain variable or value without risking erroneous typing. Previous
94 Graphical User Interface
entries are deleted. Afterwards, the combo box is closed. If there are no variables or appropriate values, the
combo box remains closed.
6.11.3 Control Buttons 95
Below the parameter display, you find five buttons that comprise the following functions:
OK Click OK to execute the operator or procedure call with the specified parameters. The execution mode depends
on the position of the PC: If the PC is placed above the insertion position, the system executes the program
from the PC until the insertion position first. Then, the operator or procedure call that has been edited in the
operator window is executed. The reason for this is that the parameter values that are used as input values
for the currently edited operator or procedure call have to be calculated. If the PC is placed at or after the
insertion position, only the currently edited operator or procedure call is executed.
The operator or procedure call is entered into the program window before it is executed. After the execution,
the PC is positioned on the next executable program line after the edited operator or procedure call.
The computed output parameter values are displayed in the variable window. Iconic variables are shown
in the current graphics window if you did not disable this option (compare section section 6.16.14 on page
125). Afterwards, the operator window is cleared. If you did not specify all parameters or if you used wrong
values, an error dialog is raised and the execution is canceled. The operator window then remains open to
allow appropriate changes.
Enter / Replace Click Enter and the currently edited operator or procedure call is transferred into the program
window without being executed. When editing existing program lines (through double-clicking in the pro-
gram window, see section 6.17.2 on page 128), the button label changes to Replace. When clicked, the
original program line is replaced.
Apply Click Apply, the operator is executed with the specified parameters, but not entered into or changed in
the program. This enables you to determine the optimum parameters rapidly since the operator dialog
remains open, and hence you can change parameters quickly. Note that this functionality is not available for
procedure calls.
Unlike the button OK, only the single line you edit or enter is executed, no matter where the PC is located.
Thus, you have to ensure that all the input variables contain meaningful values. By clicking Apply, the
corresponding output variables are changed or created, if necessary, to allow you to inspect their values.
Cancel Click Cancel to clear the contents of the operator window. Thus, there are neither changes in the program
nor in any variables.
Help Click Help to invoke the online help for the selected operator or procedure. For this the system activates the
online help window (see Help Window).
GUI Reference
6.12 Output Console Window
The output console window contains a log of the most recent messages. This includes all messages displayed on
the status bar as well as HALCON low-level errors. Therefore, HALCON low-level errors can now be logged
without an interruption of the program execution.
The window can also be opened by double-clicking the message area of the status bar. See figure 6.26 for an
example.
The logged messages are grouped into the following categories:
(Info) Info events related to the program execution, the just-in-time (JIT (page 48)) compiler, and file opera-
tions.
(Warning) Warning events related to the program execution, the just-in-time (JIT (page 48)) compiler, file
operations, HALCON low-level errors, and invalid program lines.
(Error) Program execution errors, HALCON errors, and internal errors.
(Exception) Exception events.
HDevelop uses plot windows for the visualization of data in several areas. Differences specific to a given task are
described in the corresponding sections listed below:
Plot area
x-axis
Axis limits
y-axis
The following mouse-based operations are supported to manipulate the plot display:
Operation Description
Mouse wheel on plot area Zoom in and out.
GUI Reference
Shift -drag
on plot area Zoom into a rectangular region.
Ctrl +mouse wheel Zoom x-axis.
Shift +mouse
wheel Zoom y-axis.
Drag plot area Select the visible part of the graph.
Drag on an axis Restrict the movement to the corresponding axis.
Mouse wheel on an axis Zoom in and out on the corresponding axis only.
Drag
min/max (4) on an axis Modify the displayed limits of an axis.
Ctrl Hide values temporarily while held.
adaptive The displayed data range is automatically adjusted depending on the displayed data.
increasing The displayed data range is allowed to grow if new data becomes available.
user-defined The displayed data range is never adjusted automatically, even when the data changes drastically.
The horizontal and vertical scaling can be set to linear or logarithmic mode independently.
98 Graphical User Interface
The following context menu entries are valid for all plot windows. Depending on the plot window type, additional
entries can be available which are described in the corresponding sections.
X-Axis
Action Shortcut
Description
Reset Bounds w or Ctrl+Shift+W Set width to default.
Linear Scale Set x-axis to linear scale.
Logarithmic Set x-axis to logarithmic scale.
Scale
User-defined Freeze x range
Range
Increasing Range Let x range grow on demand.
Adaptive Range Let x range grow or shrink on demand.
Y-Axis
Action Shortcut
Description
Reset Bounds h or Ctrl+Shift+H Set height to default.
Linear Scale Set y-axis to linear scale.
Logarithmic Set y-axis to logarithmic scale.
Scale
User-defined Freeze y range.
Range
Increasing Range Let y range grow on demand.
Adaptive Range Let y range grow or shrink on demand.
6.13.2 Gray Histogram Window 99
Plot Area
Action Shortcut
Description
Reset Bounds r or Ctrl+Shift+R Reset width and height to default.
Zoom In Mode + or Ctrl+Shift++ Zoom in (both axes).
Zoom Out Mode - or Ctrl+Shift+- Zoom out (both axes).
Enter Bounds... b or Ctrl+Shift+B Enter vertical and horizontal bounds para-
metrically (see above (page 97)).
Show Mouse p or Ctrl+Shift+P Visualize the mouse position with a cross
Position hair.
Show Function x or Ctrl+Shift+X Display the function value at horizontal
Value At X mouse position.
Show Function y or Ctrl+Shift+Y Display the function value at vertical
Value At Y mouse position.
Show Background g or Ctrl+Shift+G Display grid lines in the background of the
Grid plot area.
Insert Plot Code Ctrl+Shift+V Generate code to plot the current view in a
for Graphics graphics window.
Window
The gray histogram window is a tool for the inspection of gray value histograms, which can also be used to select
thresholds interactively and to set the range of displayed gray values dynamically. See also: Menu Visualization
. Gray Histogram. When opening the tool, the histogram of the image shown in the active graphics window is
displayed. When the tool is already open, the following means of sending new image data to the tool are available:
• Make another graphics window active or display another image in the active graphics window. Whenever
you do so, the histogram of this image is computed and drawn, and the tool records the graphics window
from which the image was sent (the originating window).
• Select a graphics window number in Input and Output (see below).
• Whenever image data is displayed overlaid with region data in a graphics window, you can click into any of
the segmented regions, and the histogram of the image within that region will be computed and shown. If
you click into a part of the image that is not contained in any of the overlaid regions, the histogram of the
entire image will be displayed.
GUI Reference
• The same mechanism is used for regions that have gray value information, for example, image objects created
by reduce_domain or add_channels. Here, the histogram of the image object you click into will be
displayed.
y-axis
Plot area
x-axis
Whenever new image data is evaluated in the gray histogram window, the adaptation of these values depends on
the selected adaptation mode, which can be set independently for horizontal and vertical ranges:
• adaptive
In this mode, the upper and lower boundary of the displayed gray values will always be adapted when a new
image is displayed. The maximum and minimum value for the threshold bars (green and red) are also fixed
to the maximum gray value of the type of image currently displayed.
Note that if you are using 8-bit and 16-bit images in a mixed mode, the histogram will constantly be reset. It
is not possible to display a 16-bit image, set thresholds, then display an 8-bit image and keep the threshold
values of the 16-bit image.
In adaptive mode, the displayed data range depends on the image type:
• int1, byte, direction, and cyclic images automatically use their maximum data range (for example, byte
images use the range 0 to 255).
• int2, uint2, int4, and int8 images use limits that are based on the actual data range but rounded to
powers of two.
• int2 and uint2 images can be forced to an explicit bit range using the set_system parameter
“int2_bits”.
• real images use limits that are based on the actual data range, but rounded to values that are round in
base 10 (for example, 0.1, 0.2, 0.5, or 100, 200, 500).
• increasing
In this mode, only the upper boundary of the displayed gray values will be adapted and it will only increase,
but never decrease. This for instance is useful when first inspecting 8-bit images, but then switching to 16-bit
images. In this situation, the histogram will display the 16-bit gray value range after displaying the first
16-bit image.
In this mode, the minimum and maximum value of the threshold bars are not limited to the currently displayed
image type. The reason is simple: This mode allows you to inspect images of a different data type with the
same threshold values. If the values were always limited, the histogram would “forget” the values like in the
adaptive mode.
6.13.2 Gray Histogram Window 101
• user-defined
In this mode, the boundaries are not adapted automatically (but can be changed manually). This mode is also
suitable for scenarios with images of mixed data types.
Like in the mode increasing, the minimum and maximum value of the threshold bars are not limited to the
currently displayed image type.
Histogram Options
These controls define the visible area of the histogram and the way it is displayed.
• Quantization: Display the histogram quantized. The bucket size can be specified with the slider or entered
into the spinner box. Click Auto Select to let HDevelop select a suitable bucket size automatically.
• Smoothing: Display the histogram smoothed. The smoothing factor can be specified with the slider or
entered into the spinner box. The check box specifies whether smoothing is applied or not.
Input Window specifies the graphics window of which the gray value histogram is displayed.
GUI Reference
Output Window specified the graphics window that is used for the visualization of threshold or scale operations
(see below). The visualization style is specified with the following settings:
Sometimes, it is desirable to suppress the updating of the histogram when new image data is available, for example,
if you want to select thresholds for a gradient image, but want to visualize the original image along with the
segmentation (see below). In that case you can freeze the histogram by unchecking Input Window. The currently
displayed histogram is preserved until Input Window is checked again in which case the histogram will be re-
calculated from the selected graphics window.
102 Graphical User Interface
The common context menu entries are described in section 6.13.1.3 on page 98.
Plot Area
Action Shortcut
Description
Fit Data Range d or Ctrl+Shift+D Set the lower and upper bounds so that the
histogram is displayed in its entirety.
Zoom To s or Ctrl+Shift+S Zoom to the range between the green and
Selection red line.
No output Do not process the selected gray values.
Highlight Highlight the selected gray values.
Selection
Scale Selection Scale the selected gray values.
InsertCode Generate code for the selected operation.
X-Axis
Note that histogram quantization is still performed using equal-sized linear bins if the x-axis is set to logarithmic
scale.
Y-Axis
In logarithmic mode no negative values are permitted when dragging the plot area using the mouse.
The selected range of gray values can be used for two major purposes: Thresholding (segmentation) and scaling
the gray values. This is illustrated using the images rings/mixed_03.png, and meningg5.png.
no scale scale
The setting of Output Window specifies the graphics window that is used to visualize the gray values between the
green line and the red line: You can select the originating window (Input), the active graphics window (Active),
or an arbitrary window ID from the list.
The type of visualization is specified in the table below the histogram. Click the + button to add a new operation
to the table. Click the - button to remove an operation from the table. The column Operation specifies the
6.13.2 Gray Histogram Window 103
operation that is applied to a selected range of gray values (threshold or scale, see below). To visualize a specific
operation, click the corresponding icon next to the operation .
When a multi-channel image, for example, a RGB color image, is sent to the tool, by default the histogram of the
first channel is displayed. The column Channel lets you select the channel from which to compute the histogram.
For RGB images, Channel can also be set to the special mode RGB which shows a combined histogram of all
three channels.
The columns Min and Max correspond to the position of the green and the red line, respectively. Each operation
can specify its own range of gray values.
+ button
- button
Icon
The image from which the histogram was computed is segmented with a threshold operation using the selected
minimum and maximum gray value.
With the three combo boxes Color, Draw, and Line Width in the Input and Output section of the window you
can specify how the segmentation results are displayed (see also Colored (page 56), Draw (page 56), and Line
Width (page 58)).
If you want to select threshold parameters for a single image, display the image in the active graphics window
and open the histogram tool. For optimum visualization of the segmentation results, set the visualization color to
a color different from black or white. Set Operation to Threshold and interactively drag the two vertical bars
until you achieve the desired segmentation result. The parameters of the threshold operation can now be read off
the two vertical lines.
There are two possibilities to select threshold parameters displaying the segmentation for an original image, derived
GUI Reference
from another image.
You can either display the derived image, open the histogram tool, deselect Input Window, display the original
image, and then select the appropriate thresholds. This way, only one window is needed for the visualization.
Or, you can display the derived image in one graphics window and the original image in another. Activate the first
graphics window image, and make sure Input Window is checked so that the corresponding gray value histogram
is calculated. Afterwards, Input Window can be turned off to prevent the histogram from being updated. In the
gray histogram window set Output Window to the window ID of the second graphics window and select your
thresholds.
Multiple Threshold Operations
If multiple operations are visualized at the same time, the display depends on the combo box below the table of
operations:
If none is selected, the results of the different threshold operations are displayed independently.
If union is selected, the results are combined to a single region.
If intersection is selected, only the common pixels from all results are visualized.
Connected Regions
Clicking Connection displays the connected regions of the selected gray values in the style specified with Color,
Draw, and Line Width.
This display mode is similar to a plain threshold operation. Additionally, it performs a connection operation.
The separate regions can only be distinguished if Color is set to colored 3, colored 6, or colored 12.
104 Graphical User Interface
Click the button Insert Code to generate HDevelop code that performs the currently visualized threshold opera-
tion(s) in your program. The code is inserted at the IC.
The resulting regions of the threshold (and connection) operation can be used as input to the feature histogram
window or the feature inspection window if the gray histogram window is kept open. These windows are described
in the next sections.
The scale operation maps the gray values between the green line and the red line to the full range (usually 0...255).
See also scale_image.
The gray values of the image are scaled such that the gray value 0 of the scaled image corresponds to the selected
minimum gray value and the gray value 255 to the selected maximum gray value. The combo box Output Window
determines the graphics window, in which the result is displayed. Use this to interactively set a “window” of gray
values that should be displayed with a large dynamic range. Note that, not more than one scale operations can be
visualized in the graphics window at the same time.
Click Insert Code to generate HDevelop code that performs the currently visualized scale operation in your
program. The code is inserted at the IC.
Most parts of the tool are built up similarly to the gray value histogram, which is described in detail in section 6.13.2
on page 99 (Menu Visualization . Gray Histogram). Reading this description beforehand is highly recom-
mended. In the following, we concentrate on points specific to the feature histogram. An important point regards
the “source” of the regions or XLDs: The feature histogram is calculated for the regions or XLDs that were dis-
played most recently in the graphics window. Thus, if you display an image, and there are no regions or XLDs,
the histogram remains “empty”. As soon as you display regions or XLDs on top of an image, the histogram is
calculated. If you display regions or XLDs without an image, you can still calculate feature histograms, but only
for shape features. Please keep in mind that only the most recently displayed regions or XLDs are the source of
the histogram, not all objects currently displayed in the graphics window.
The histogram itself is displayed with the horizontal axis corresponding to the feature values and the vertical axis
corresponding to the frequency of the values, for example, to the number of regions or XLDs with a certain feature
value.
When comparing feature histograms to gray value histograms, you will note a typical difference: Because in most
cases the overall number of regions or XLDs is much smaller than the overall number of pixels, feature histograms
often consist of individual lines, most of them having the height 1. Of course, this effect depends on the selected
6.13.4 Feature Inspection Window 105
feature: For features with floating-point values, for example, the orientation, the probability that two regions or
XLDs have the same feature value is very small, in contrast to features with integer values, for example, the number
of holes.
You can influence the calculation of the histogram with the slider Quantization. The selected value is used
to discretize the horizontal axis: Instead of determining the frequency of an “exact” feature value, regions with
feature values falling within discrete intervals are summed. Graphically speaking, the horizontal axis is subdivided
into “bins” with a width equal to the value selected with the slider Quantization.
GUI Reference
+ button
- button
Icon next to the operation
You can add additional features using the + button , or remove features using the - button . As with the gray
histogram operations, each selected feature has to be enabled to visualize the selection in the graphics window.
Select
Selected region
Feature value of selected region
Range visualization
The strategy to determine the data from which to compute the features is very similar to that of the gray histogram
inspection window (see section 6.13.2 on page 99). You can display an image or region by double-clicking it in
the variable window or you can select a region or an image which is already displayed by single-clicking it. If
you display or click into an image, the gray value features of the entire image will be calculated. If you click
into a region that is not underlaid with an image, only the shape features of this region will be displayed. If you
click into a region that is underlaid with an image or into a region that has gray value information (for example,
from reduce_domain or add_channels), both the shape and gray value features of that region will be displayed.
Finally, if you have overlaid an image with a region, but click into a part of the image that is outside the region,
only the gray value features of the entire image will be calculated.
Use the “select” tool of the graphics window to select a region or XLD. The selected region or XLD is highlighted
in the graphics window. The corresponding variable name and index are displayed in the title of the feature
inspection window.
The gray value features of a multi-channel image are calculated from all channels independently.
The tree on the left side of the feature inspection window groups the features into categories.
• Region features: This group contains features that describe the selected region, for example, area, center, and
orientation.
• Gray value features: The feature values of this group are calculated from the gray values of the image under
the selected region, for example, minimum and maximum gray value, mean gray value, anisotropy and
entropy.
• XLD features: This group contains features that describe the selected XLD (for example, its dimensions or
shape properties).
You can select the features to be inspected by ticking the corresponding check boxes in the tree. The selected
features are displayed on the right side of the window. For each feature the calculated value of the selected region
or XLD is displayed (or the value for the entire image). The current value is also visualized as a gauge in a value
range that can be set to the desired values. Select Show Minimum/Maximum, which is available in the context menu
of the right side of the window.
See figure 6.31 for an illustration of a feature inspection. The range for the area feature has been set to [2000,
7000]. Individual rings can be inspected by selecting them in the graphics window.
6.13.5 Line Profile Window 107
Moving the mouse pointer over a feature value displays a tool tip. It shows the name and short description of the
HALCON operator used for the calculation of that value. Using the context menu, you can insert the corresponding
operator into the operator window.
GUI Reference
When opening the line profile without using the Measure Assistant, nothing is displayed in the line profile until an
ROI is drawn in the graphics window using the corresponding buttons in the line profile window.
108 Graphical User Interface
There are various options for linking the line profile to different windows:
• Make another graphics window active, select another window ID in the Input Window drop-down menu, or
display another image in the active graphics window. Whenever you do so, the line profile of the ROI in this
image is computed and visualized, and the tool stores the graphics window from which the image was sent
(the originating window).
• A Measure Assistant can be selected as data source by activating the checkbox Measure Assistant under
Input Window within the line profile window and choosing the correct assistant from the drop-down menu
if several assistants are open at the same time.
• The line profile can be opened from the Measure Assistant window by clicking on the View Line Profile
button on the Edges tab.
If the Measure Assistant (page 219) is not selected as source of an ROI, a new ROI can be created and edited
using the ROI menu buttons above the line profile display. Those buttons allow you to draw either a linear or a
circular ROI, delete all ROIs or view the ROI shape. Once an ROI has been created with the Draw Line or Draw
Circular Arc buttons, those ROI creation buttons are grayed out because the line profile window can only handle
one ROI at a time.
The main part of the tool is the area in which the gray-value profile of the image along an ROI is displayed in blue.
The horizontal axis represents the length of the ROI in pixels and therefore gives the position of the gray values
along the ROI.
The dynamic parts of the line profile area are the two colored lines, which can be manipulated. The vertical green
and red lines denote the minimum and maximum selected position along the ROI, respectively. Those lines are
also displayed in the active graphics window. Their visualization can be adapted under Output. Another dynamic
part is the vertical coordinate axis, which displays the gray values in the image. For byte images, this ranges from
0 to 255. As it comprises only the gray-value range between darkest and the brightest pixel, these values do not
usually start with 0 and end with 255. For all other image types, for example, real images, the horizontal axis runs
from the minimum to the maximum gray value of the image and the labeling of the axis is changed accordingly.
Initially, the line profile is displayed at full vertical range, for example, up to the peak value. With the buttons to
the left of the line profile display, you can modify the displayed part:
The common interaction with plot windows is described in section 6.13 on page 96.
The common context menu entries are described in section 6.13.1.3 on page 98. The context menu entries specific
to the line profile window are listed below:
Plot Area
Action Shortcut
Description
Fit Data Range d or Ctrl+Shift+D Set the lower and upper bounds so that the
line profile is displayed in its entirety.
Zoom To s or Ctrl+Shift+S Zoom to the range between the green and
Selection red line.
6.13.5 Line Profile Window 109
Smoothing: Display the profile smoothed. You can smooth the profile before displaying it by specifying a smooth-
ing factor with the slider in the spinner box and by clicking the checkbox. If smoothing has been applied with the
Measure Assistant (page 219), those values are automatically adopted. The smoothing can then be performed in
the line profile window as well and likewise affects the Measure Assistant’s smoothing (see figure 6.33).
Derivative: Display the derivation of the line profile.
GUI Reference
Force minimum line profile width: Reduce the line profile visualization to a minimum width by using the
button .
This section lets you choose the source of your image and therefore the source of the ROI. By default, the active
graphics window is chosen into which an ROI can be drawn using the ROI menu of the line profile window. The
110 Graphical User Interface
ID of another graphics window can be selected from the drop-down menu. If a Measure Assistant (page 219) is
activated, one assistant can be chosen from the drop down menu or it can be opened by activating the Use Measure
Assistant as Data Source button.
The remaining buttons let you choose the visualization of the marker lines the line profile displays within in your
active ROI. You can select the color and length of your lines as well as their width. Changing those output features
can be necessary to achieve optimum visibility within an image.
6.13.5.6 Statistics
The Statistics section of the line profile window displays the values from the line profile display above. It
includes values for the position (x Value) as well as the gray values. It therefore informs you where interesting
gray values can be found. Those gray values include
• the Peak, which marks the position of the highest gray value,
• the Range, defining the length of the ROI as well as the range of gray values along the ROI.
• Value at Min and Value at Max, defining the position and gray value of the above determined selection
(that was localized with the green and red line) as can also be seen in the active graphics window.
• There are two further values that only concern the gray values: Mean defines the mean gray value while
Deviation is the average deviation.
When focusing your camera, it might help to take a quick look at the gray-value transitions along a line within the
image to see whether the edges are sharp or still a bit blurry. While sharp images are defined by abrupt changes
between dark and bright gray values, no abrupt changes but rather gray-value transitions can be found in blurry
images (see figure 6.34).
a)
b)
The tab card General displays file properties of the current program, such as file name, path, creation and modi-
fication date, and write permission. It also shows the file size, the number of lines of code, used and unused local
procedures, used external procedures and used protected procedures. This is displayed in figure 6.35.
The tab card Used Modules lists the HALCON modules used by the current program. Modules marked with
are used. Modules marked with are potentially in use by dynamic licensing. This window allows you to get
an estimate of how many modules your application will need in a runtime license. Please refer to the Installation
Guide, section 5.4.2 on page 29 for more information about modules, runtime licenses, and dynamic licensing. See
figure 6.35 for the corresponding dialog of a 3D object processing example.
Check only used procedures If checked, only used procedures are considered for the evaluation of the used
modules. Otherwise, all procedures are considered.
GUI Reference
Copy to Clipboard Copy the names of the used modules to the system clipboard. This way the list can be
easily pasted into other applications.
Procedure Options
The bodies of locked procedures (see section 5.6 on page 47) are not printed.
Windows: %APPDATA%\MVTec\HDevelop.ini
Linux: $HOME/.hdevelop/MVTec/HDevelop.ini
HDevelop can read additional preferences from another startup file specified with the command line switch
-add_preferences <file>. To load the preferences only from another startup file, start HDevelop with
the command line switch -load_preferences <file>. In both cases the settings will be stored back to
HDevelop.ini when HDevelop is quit. To use a different startup file altogether, start HDevelop with the command
line switch -use_preferences <file>.
Export Using the check boxes you can specify which settings will be saved to the selected file.
Import In the import dialog, you can select a file with previously saved HDevelop preferences (default file exten-
sion: .hdp). This selection of preferences was saved using the menu entry Export. The check boxes allow
you to import groups of settings selectively. They correspond to the categories of the dialog. The runtime
settings are not persistent and can neither be exported nor imported.
Reset Selecting this menu entry resets all preferences (except the window geometry and layout) to the default
settings. If you want to reset the window geometry as well you can start HDevelop with the following
command line switch:
hdevelop -reset_preferences
The preferences dialog contains a list of categories on the left and several related tab cards on the right. The size
of these elements is controlled by a splitter. The available categories of preferences are described in the following
sections.
6.16.1 User Interface . Program Window 113
Font
Specifies the font that is used in the program window.
Indent Size: Specifies the number of spaces an indenting level in the program window accounts for. In HDe-
velop, the bodies of loops and conditionals are indented automatically.
Editor Settings: These settings specify the editing behavior in the program window.
ToolTip Mode:
When the mouse cursor rests on a program line, a tool tip is displayed by default. When placed over an
operator name, the interface of the corresponding operator and its short description are displayed. When
placed over an instantiated iconic variable, an icon is displayed. When placed over the program line column,
warning messages may be displayed, for example, if the used operator is deprecated. This behavior can be
GUI Reference
turned off if it disturbs your editing activities.
• Always
Always display a tool tip when the mouse cursor rests on a program line.
• Only on Control key
Display the tool tip only when the Ctrl key is pressed.
Autocompletion:
• Advanced
Enables advanced autocompletion in the program window (see section 6.17.2 on page 128).
• show after [ms]:
This value defines the delay before the autocompletion is displayed.
Auto Highlight:
Enables the auto highlight feature which can highlight
• occurrences of the word under the cursor,
• occurrences of the currently selected string,
• matching brackets, and
114 Graphical User Interface
Whitespaces
Enables the visualization of whitespaces within the program window. By default this option is turned off.
• General
The font used throughout the user interface (menu entries, labels etc.)
• Help Window
The body font used in the help window (menu Help (page 64) . Help).
• Program Window
The font used in the program window. This is the same font setting as on the tab card Program Window (see
above).
• Advanced Autocompletion
The font used in the advanced autocompletion overlays.
• Values and Parameters
The font used for displaying values in the variable window and associated inspection windows as well as
parameters in the operator window.
• Printing
The font used when printing program listings.
By default, HDevelop uses the language that is specified as the display language of the operating system. If the
language is changed via these dialogs, the operating system locale is ignored.
• HDevelop GUI
Changes the language of HDevelop’s Graphical User Interface and the language of the Program and Proce-
dure Reference within HDevelop’s help window. The new language will be applied the next time HDevelop
is started.
• Operator Reference
Changes the language of the HALCON Operator Reference within HDevelop’s help window (page 81), as
well as the language of the operator descriptions within the autocompletion and the operator window (page
91).
The new language of the operator descriptions is applied immediately. Note that some changes, like the
language of the operator window, only take effect after the next time HDevelop is started.
6.16.4 User Interface . Themes 115
• Current theme
Specifies the appearance of HDevelop, for example, the background color and the icon set. You can choose
between a dark and a light theme.
• Colors
Specifies the colors of the current theme. Click the color field to select a color. To reset a color, click . To
reset all colors, click Reset all colors.
Use this tab card to manage the list of directories that contain external procedures. The directories are recursively
scanned for external procedures in their listing order (see section 5.4 on page 45).
GUI Reference
Figure 6.38: Procedures . Directories.
Libraries: For each directory, the number of libraries is displayed. The number in parentheses sums up the
number of procedures contained in libraries.
Procedures: This column displays the total number of procedures, that is, the number of procedures found in
libraries and all other external procedures.
Loaded: The number of loaded procedures is usually equal to the total number unless the directory contains
corrupted procedure files, or files with the wrong permissions.
Procedures with the extension .hdvp always override procedures with the same name but the extension .dvp in
the same directory.
Please note that HALCON comes supplied with a set of standard procedures. These are general-purpose procedures
used by many of the supplied example programs. The path cannot be altered or deleted. It is, however, possible
to override the supplied external procedures by placing external procedures with the same name in one of the
user-defined directories.
The documentation of the supplied procedures is available in the online help of HDevelop under Procedure Refer-
ence Manual.
116 Graphical User Interface
Add: Select an additional directory from the file selection dialog. This directory will be added to the list. All
subdirectories of the selected directory will be scanned as well.
Add Library: Explicitly add a library (.hdpl) file to the list of directories. This will make the library procedures
available without looking at other procedures in the same directory.
Delete: Delete the selected entry from the list. Programs using any external procedure from that directory will
no longer run.
Rescan: Rescan all listed entries to reflect any changes in the file system.
Using the check boxes you can disable the corresponding directories temporarily without removing them from the
list.
The list entries can be reordered using the up/down arrow buttons.
This tab card lists all external procedures in the order they are loaded from the configured directories (page 115).
For each procedure, the following information is displayed:
Column Meaning
# Number of the external procedure.
Activated The check box toggles the activation status of the corresponding procedure, that
is, whether the procedure can be resolved. It can be used to temporarily dis-
able procedures, for example, to make concealed procedures of the same name
available.
Procedure Name Name of the external procedure.
State Loaded: The external procedure has been loaded successfully. It can be used in
any HDevelop program.
Unloaded: An error occurred while trying to load the external procedure, for ex-
ample, the file permissions are wrong or the external procedure file is corrupted.
Search Directory Directory name from the tab card Directories where this procedure is found.
Relative Path Path name of the external procedure relative to the search directory.
Used by Usage counter and the names of the callers of this procedure.
Modifications The number of modifications to the external procedure after it has been loaded.
Using the context menu, basic file system operations can be executed. You can copy, delete, or rename the selected
procedure file in the file system.
The table on the left lists all libraries, the number of contained procedures, and the path name to the library file.
The special entry “Procedures” contains all local procedures and all external procedures that are not contained in a
library.
GUI Reference
The following buttons are available:
For more information about procedures and their properties, see section 5.1 on page 43 and section 5.6 on page 47.
Drag-and-drop support:
In practice, it is much easier to drag procedures from one location to another than using the corresponding buttons.
A simple drag-and-drop moves the selected procedures from one place to another. To copy the selected procedures,
start to drag, then hold down Ctrl , and drop at the destination (this is visualized by a + icon).
118 Graphical User Interface
Using this tab card, you can conveniently manage the status and passwords of all procedures (local and external).
The procedures are divided into three categories (from left to right): Procedures without a password (unprotected),
procedures for which the password has already been entered in this session (unlocked protected), and procedures
that are locked with a password (locked protected). For an explanation of the different states, see section 5.6 on
page 47.
Using the arrow buttons between the columns or the left and right cursor key, you can move the selected procedures
to a different status. If you move procedures from the first to the second column, a password dialog is displayed
which is described in section 6.17.9 on page 146. The same password is applied to all selected procedures.
GUI Reference
If you move procedures from the second to the third column, the corresponding procedures will be locked. The
procedure interfaces can no longer be edited and the procedure bodies will no longer be displayed. They can
only be accessed if the correct password is supplied. This can either be done from this dialog by moving the
corresponding procedures back to the middle column and entering the password. Or, you can unlock procedures
individually from the program window as described in section 6.17.9 on page 147.
If you select multiple procedures in the third column and move them to the left, a password dialog appears to
unlock the procedures. Only those procedures are moved (and thus unlocked) that match the supplied password.
This way, you can conveniently edit a group of procedures that share the same password.
If all local procedures of the current program have been protected at once (by protecting the main procedure and
enabling the corresponding option, see figure 6.60 on page 147, the local procedures will always move as a single
group, even if only one local procedure is selected. The icon next to Program Protection State: (at the
bottom of the dialog) is marked with a lock if the entire program (main procedure and all local procedures) is
protected.
The button Change Password is available if one or more procedures are selected in the middle column. It assigns
a new password to the selected procedures, regardless if the previous passwords were different.
Please note, that password changes or moving procedures from or to the first column require the corresponding
procedures to be saved. See Save (page 53) and Save All (page 53).
120 Graphical User Interface
This tab card lists the usage of procedures grouped by their calling procedures. You can select a procedure and the
type of used procedures (either local or external). For the main procedure you can also list the unused procedures.
The tab card is displayed in figure 6.43.
This tab card helps you to find unresolved procedures in your current program. If the current program or the loaded
procedures contain unresolved procedure calls, they are listed here along with the calling procedure name.
6.16.11 General Options . General Options 121
GUI Reference
Variables can be sorted by name or by occurrence. See also variable window (page 164).
• Return relative file paths when appropriate
If enabled, HDevelop will try to convert file names selected from file selection dialogs to relative path names.
HALCON uses several distinguished directories that act as the preferred search directories for certain oper-
ators. For example, read_image will at first search for image files in %HALCONROOT%/images when given
a relative file path. For such operators HDevelop tries to turn the selected path into a path relative to the
preferred search directory.
If this fails, HDevelop tries to turn the selected path into a path relative to the current working directory. As
a minimum requirement, the selected path and the current working directory must have the same device and
top level directory. This will be useful in cases where, for example, an image directory and a directory for
HDevelop scripts are placed side by side within a common project directory.
If no relative path can be established, HDevelop will use the absolute path.
• Replace backslashes by slashes in pasted paths (Windows only)
This option determines the behavior when Windows path names are copied from an external program and
pasted into HDevelop. If set to Always, backslashes will always be replaced by slashes. If set to Never, the
path will be pasted unaltered. If set to Ask, HDevelop lets you choose the behavior each time a Windows
path is pasted from the clipboard. For relative path names the replacement will only be performed if the
path is pasted to a position where a path name is expected, for example, an operator parameter with the
corresponding semantic type.
122 Graphical User Interface
• Copy local procedures in full-text editor This option determines the behavior when program
lines containing calls to local procedures are copied and pasted. If set to Always, the corresponding local
procedures will be copied as well unless they are already available. If set to Never, the corresponding local
procedures will not be copied. Note that this will result in invalid calls if the copied lines are pasted into a
program where the corresponding local procedures are not available. If set to Ask, HDevelop lets you choose
the behavior each time you paste the lines into a program where the local procedures are not available.
• Number of recent program files
The maximum number of recent program files that are displayed in the menu File . Recent Programs
(page 53) can be adjusted by altering this value.
• Show recent program files
If you select the option Only available, the list of recent programs contains only programs that are cur-
rently available. This option is useful, if the list is likely to contain files from network drives that might be
disconnected at times.
• Show full path in main window title
If enabled, the full path of the current program is displayed in the title bar of the HDevelop window. Other-
wise the file name is displayed.
• Load the latest program automatically while starting HDevelop
If enabled, the most recently used program, that is the top entry under File . Recent Programs (page
53), is loaded automatically when HDevelop starts. This behavior can be overridden from the command line
using the switches load_recent_prog/load_no_prog (see appendix C.1 on page 323).
• Save program and external procedures automatically before execution
If this option is enabled and you click any of the execution buttons (like Run or Step Over) and there are
unsaved changes in the current program, the program will be saved before being executed. Use this option
with care: You usually do not want to select this option if you are experimenting with a program, for example,
when trying out different parameter settings.
• Encoding for saving HDevelop programs and procedures when the legacy HDevelop 5.0
- 9.0 file format (dev, dvp) is used
HDevelop supports two different file types for programs and procedures (see section 5.2 on page 44). Pro-
grams and procedures in the default file format (.hdev and .hdvp, respectively) are always saved in UTF-8
encoding. When saving programs and procedures in the legacy format (.dev and .dvp, respectively), the en-
coding depends on this setting. In order to exchange data with older versions of HDevelop (before HALCON
8), the encoding must be set to Native.
• Default file format for saving HDevelop programs
Sets the default file format for new programs, see section 5.2 on page 44.
• Default file format for saving HDevelop procedures
Sets the default file format for new external procedures, see section 5.2 on page 44.
• Update windows in single-step mode
Specifies the window update policy when stepping through the program. The different options are described
in section 6.1.4 on page 58.
• Maximum number of subthreads
Specifies the maximum number of subthreads that be started using the par_start qualifier (see section 8.11
on page 283).
• Window open offset
Specifies an offset for the origin of floating windows. These offset values are added to the values for Row
and Column, which are input parameters for some HALCON operators, like dev_set_window_extents
or dev_open_tool. For example, dev_open_tool(’bookmarks_dialog’, 0, 50, ...) with offset
values (50, 50) opens at position (50, 100). See figure 6.44 for an illustration of the window offset.
• Origin of coordinates
Sets the reference point for the origin of floating windows, opened via operators like dev_open_tool and
dev_open_window.
6.16.12 General Options . Experienced User 123
Reference point for the window position by default (upper left corner of the desktop)
The origin of the HDevelop window is its upper left corner
Value definded by the operator values for Row here (0, 0)
Value definded by the operator values for Column here (0, 0)
• Main display
The main display is usually defined in the display settings of your operating system. If you have only
one monitor, this is the main display. This display acts as the active desktop. It shows the taskbar i
and most items will open on the main display by default.
• Display showing HDevelop
The display that currently shows the HDevelop main window (page 51). This setting is only relevant
if you use multiple monitors. Otherwise, the Display showing HDevelop is the same as the Main i
display.
• HDevelop
HDevelop’s main window (page 51).
Please also note the setting Window open offset.
GUI Reference
• Floating windows behavior
Specifies the accessibility of floating windows. To apply the new settings, a restart of HDevelop is required.
This only refers to graphic, tool and assistant windows.
• Non minimizable and remains in foreground
Floating windows stay in the foreground of HDevelop. They cannot be minimized, nor can they be
accessed separately via taskbar.
• Minimizable, does not remain in foreground
Floating windows move to the background as soon as an other window
gets the focus. They can be
minimized and get a taskbar entry. Thus, you can access them via Alt+Tab under Windows. (Depending
on your operating system, the shortcut for quickly switching between windows might differ.)
HDevelop normally displays a dialog when a run-time error occurs (unless this has been changed in the
tab Runtime Settings . Runtime Settings). Per default, these dialogs are suppressed when the error
occurs in a watched block of program lines (surrounded by try ... catch). In this case, the exception is
caught by the program. The program continues at the corresponding catch ... endtry block.
Deactivating this option will bring up a dialog even if the error occurs within a try-catch block.
Error message dialogs within try-catch blocks are always suppressed in JIT-compiled procedures (see
option below) regardless of this setting.
• Stop execution at invalid program lines
If enabled, a dialog will be displayed if an invalid program line is reached, allowing to stop the program and
edit the corresponding line, or to ignore it and continue execution. If disabled, invalid program lines will
always be ignored.
• Suppress warnings for HALCON operators that are replaced by dev-operators
Some operators are deprecated in HDevelop, and issue a warning message when selected in the operator
window. Activating this option suppresses the warning message.
• Disable parameter detection for acquisition operators
In the operator window, the parameter suggestions for the operators open_framegrabber,
set_framegrabber_param, and get_framegrabber_param depend on the selected image acquisi-
tion interface. This behavior can be disabled by ticking the check box. See also section 6.11.2 on page
92.
• Ignore semantic type
By default, the parameter suggestions in the operator window and the program window (with advanced
autocompletion enabled) include only variable names that match the semantic type of the corresponding
parameter. If Ignore semantic type is checked, these suggestions are extended so that they also include
suggestions of variables with a different semantic type.
• Show memory usage
If this option is activated, the internal temporary memory usage of the last operator or procedure call is
displayed in the status bar.
• Execute procedures JIT-compiled
HDevelop supports just-in-time (JIT) compilation of procedures for optimized performance of HDevelop
programs, see section 5.9 on page 48. The JIT compilation is globally enabled or disabled using this setting.
• File tracking behavior
By default, the file tracking is set to System file tracking. This option may limit the number of tracked
files. When Application file tracking is activated, a timer checks for changed files. This option can
affect the operating system performance. Switch off the file tracking by choosing No file tracking.
6.16.13 Visualization Settings 125
The tab cards in this category define the default visualization settings for graphics windows when HDevelop is
started. See section 6.8.2 on page 76. The dialog is displayed in figure 6.45.
Reset to factory defaults Remove the user-defined colors and custom color sets.
Adapt active settings Adapt the user-defined colors and custom color sets permanently. To define your own
colors and color sets, use the Visualization Parameters dialog (see section 6.8.2.1 on page 77).
The settings in this category affect the runtime behavior of HDevelop. Please note that the runtime settings are not
persistent between sessions. The runtime settings are reset to their default values, when a new HDevelop program
is started with Menu File . New Program.
• Give Error
GUI Reference
This check box specifies the behavior of HDevelop if an error occurs. If it is checked, HDevelop stops the
program execution and displays an error message. Otherwise the error is ignored. See also: dev_set_check.
• Show Processing Time
This check box indicates whether the required runtime of the last operator or procedure call should be dis-
played after the execution has stopped. It is a measurement of the needed time for the current operator or
procedure call, without output and other management tasks of HDevelop. Along with the required runtime,
the name of the operator or procedure is displayed in the status bar at the bottom of the main window. Please
note that the displayed runtime can vary considerably. This is caused by the inaccuracy of the operating
system’s time measurement procedure. See also: dev_update_time.
This option can also be toggled from the context menu of the status bar (see section 6.3 on page 65).
• Update Program Counter
This option concerns the display of the current position while running the program. The PC always indicates
the line of the currently executing operator or procedure call or the line before the next operator or procedure
call to execute. Using the PC in this way is time consuming. Therefore, you can suppress this option
after your test phase or while running a program with a lot of “small” operators inside a loop. See also:
dev_update_pc.
• Update Variables
This check box concerns the execution of a program: Every variable (iconic and control) is updated by default
in the variable window. This is very useful in the test phase, primarily to examine the values of control data,
126 Graphical User Interface
since iconic data is also displayed in the graphics window. If you want to save time while executing a program
with many operator calls, you can suppress this output. Independent of the selected mode, the display of all
variables will be updated after the program has stopped. See also: dev_update_var.
• Update Graphics Window
This item concerns the output of iconic data in the graphics window after the execution of a HALCON
operator. With the default settings, all iconic data computed in the run mode is displayed in the current
graphics window. You may want to suppress this automatic output, for example, because it slows down the
performance or because the program handles the visualization itself. If the output is suppressed, you have
the same behavior as exported C, C++, Visual Basic.NET, or C# code, where automatic output of data is not
supported. See also: dev_update_window.
• Enable the context menu in the Graphics window
If this option is activated, the context menu is available when clicking in a graphics window with the right
mouse button. This behavior can be undesirable if a program provides user interaction with the mouse. See
also: dev_set_preferences.
• Enable the mouse wheel in the Graphics window
By default the mouse wheel is used to zoom in and out in the graphics window. If this interferes with a
custom mouse handling, the mouse wheel can be disabled. This is desirable, for example, if 3D objects are
displayed in the graphics window and the zooming functionality is implemented with the help of 3D display
operators. See also: dev_set_preferences.
• Enable tooltip showing coordinates and gray value at the mouse cursor position...
If enabled, a tooltip will be displayed if the mouse cursor is in the graphics window and the Ctrl key is held
down. The tooltip shows the pixel coordinates (row and column), and the gray value(s) at the mouse cursor
position.
• Record Interactions
If enabled, changes to the visualization settings will be recorded by adding the corresponding operator calls
to the current program. See section 6.4 on page 57.
This tab card allows you to modify the default behavior of stop and wait_seconds. Click the check box next to
the corresponding operator and specify the modification.
stop: By default, the stop operator halts the program execution. It is mainly used to highlight or evaluate
processing steps in a larger program. If you want to run such a program uninterrupted without altering the
actual program code, you can make the stop operator behave like wait_seconds, that is, perform a defined
delay. The delay is specified in seconds.
wait_seconds: The operator wait_seconds is often used in situations where an intermediate program result
is presented that would otherwise pass by too fast. Sometimes, you want to run such a program without
any delays, for example, for performance measuring purposes. At other times, you would like to stretch the
delays, for example, for evaluation or presentation purposes. Therefore, you can redefine the actual delay of
wait_seconds. The delay is specified in seconds.
Selecting exactly causes the specified delay for each wait_seconds instruction.
Selecting with minimum causes delays up to the specified duration. Calls to wait_seconds with a shorter
duration will not be affected.
Selecting with maximum causes delays of at least the specified duration. Calls to wait_seconds with a
longer duration will not be affected.
The following table shows the actual delay caused by different override settings:
Actual program call with minimum(10s) with maximum(10s)
wait_seconds(1) 1s 10s
wait_seconds(20) 10s 20s
6.16.16 Telemetry 127
6.16.16 Telemetry
You can help us improve your HALCON user experience by anonymously sharing usage data with MVTec. This
will also allow you to receive additional information on our latest or upcoming products. To enable the telemetry
feature, select Share the following anonymous data with MVTec and decide which information to send:
System information The sent data comprises geographical data, such as country and region, operating system
and platform, monitor resolution, graphic card models, and the number of CPU cores. Note that this option
cannot be excluded separately.
HDevelop usage statistics The sent data comprises system information about HDevelop, such as selected
language, HALCON version and edition, workflow information, like export and tool usage.
HDevelop Help usage statistics The sent data comprises workflow information inside HDevelop Help, like
used search terms and viewed chapters and pages.
HALCON Library statistics The sent data comprises the list of used HALCON modules.
You can inspect the raw data being sent by clicking Get session data. To export the data as a CSV file, click
Export session data.
A program window displays the HDevelop program code of a single procedure. You can open multiple program
windows to view different procedures or different parts of the same procedure at the same time (see Menu Window
. Open Program Window (page 63)). To switch between selected procedures rapidly, program windows support
multiple tabs.
The program window (see figure 6.46 on page 129) is divided into three areas:
• The main part of the program window contains the program code of the current HDevelop procedure. See
section 6.17.2.
• The column at the left side displays line numbers. It also contains the PC, the IC, and optionally, one or more
breakpoints. Bookmarked lines and warning indicators are displayed. See section 6.17.3 on page 133.
• At the top, the displayed procedure can be selected from the drop-down list. If you hold down Ctrl while
GUI Reference
selecting a procedure, it will be displayed in a new tab card.
The arrow buttons provide convenient access to previously displayed procedures. For example, if the current
procedure is the main procedure, and you select another procedure from the drop-down list, the left arrow
button takes you back to the main procedure. When you get there, the right arrow button moves forward, and
displays the previously selected procedure again.
Using the tool bar button to the right of the procedure list, the interface of the current procedure and its
documentation can be edited. The number of parameters as well as their names and types, can be modified.
See section 6.17.5 on page 135 for a detailed description.
BP
Warning
PC
IC
Bookmark
Browse history
Add new tab
Select procedure
Edit procedure interface
Pin program listing
GUI Reference
Comments
Comments start with an asterisk (*) as the first character. You can also add comments at the end of regular program
lines by introducing them with //.
• Variable assignments
FileName := 'clip'
assign('clip', FileName) // deprecated
• Setting individual tuple elements
130 Graphical User Interface
Line[12] := 'text'
assign_at(12, 'text', Line) // deprecated
Note that for loops always have to be entered in the following format:
Line Continuation
Operator calls can span several lines for readability. A line can be continued by entering a backslash character as
the last character of that line.
For example, you can enter
disp_arrow (WindowID, \
Row[i], \
Column[i], \
Row[i]-Length*sin(Phi[i]), \
Column[i]+Length*cos(Phi[i]), \
4)
instead of
Auto Indenting
The indent of new lines is adjusted automatically. Usually, the indent of the previous line is maintained. If a line is
continued inside the parentheses of an operator call, the new line is indented up to the opening parenthesis. If the
previous line opens a control structure (for example, if or while), the indent is increased by the indent size. The
indent size is specified in the preferences (see Program Window (page 113)). It defaults to four spaces. If a control
structure is closed (for example, by entering endif or endwhile), the indent of the current line is decreased by
the indent size.
Advanced Autocompletion
The program window provides advanced autocompletion to support you in several ways.
• Getting started
HDevelop’s autocompletion can assist you getting started with a completely empty program. Press Tab to
get a list of typical first operators like dev_open_window and read_image.
Press Shift+Tab to get suggestions for possible predecessors.
• Typing support
When you start typing a new line, HDevelop will suggest a list of matching operator names:
Note that the line is highlighted as invalid (red in the default color scheme) because it is still incomplete.
The list is updated immediately as you continue typing:
GUI Reference
Press Tab to complete to the longest common string. In this example, only one operator name remains in
the list. Thus, it is fully completed, including the opening parenthesis of the operator call:
Once the cursor moves inside the parentheses, the suggestion list changes from operator mode, to parameter
mode. The signature of the selected operator is displayed, and the parameter corresponding to the cursor
position is highlighted in bold.
The first entry in this list is a suggestion that completes the full operator call up to the closing parentheses.
Again, typing ahead updates the list of suggestions accordingly. The remaining entries are suggestions for
the first parameter of the operator call.
Press Tab to select the first suggestion,
or press Up or Down to step through the list entries,
132 Graphical User Interface
and press Tab or Return to select the highlighted entry. Then, enter a comma or press Tab again to get
suggestions for the second parameter:
Double-clicking the browse button in the suggestion list opens up a file selection dialog to specify the file
name parameter. These browse buttons are included in the suggestions lists of all parameters that specify a
file name.
In this example, we want to load the image clip, so none of the suggestions fits. Just type the file name
in single quotes ('clip') and press Tab to complete the parameter list. The closing parenthesis is inserted
automatically:
When entering expressions, the advanced autocompletion also suggests operations supported by the HDe-
velop language (see chapter 8 on page 247).
Please note that the advanced autocompletion does not suggest legacy or deprecated operators. If you enter
such an operator manually, a warning icon is displayed in the left column. Move the mouse cursor over this
icon to get a tool tip with a corresponding warning message:
The program window also supports autocompletion of block statements. When entering a control statement,
for example, the control statement of a for loop,
the corresponding end statement, for example, endfor of
the block is automatically inserted by pressing Return .
The PC can only be placed on program lines of procedures on the callstack. For example, in figure 6.47 on page
136, the PC can be placed anywhere in the procedures main, first, or second. It may not be placed within the
procedure third. If the PC is placed in first, the first element of the callstack is popped. If the PC is placed in main,
GUI Reference
only main remains on the callstack. Please note the outlined green arrows in main and first: They indicate the
return positions.
By clicking into the program window with the right mouse button you can open a context menu, which contains
shortcuts to some of the actions of the menus Menu Edit, for example, copy and paste lines, and Menu Execute,
for example, activate and deactivate lines or set and clear breakpoints. Please note that these actions behave slightly
differently than their counterparts in the main menus: When called via the main menus, the actions are performed
only on the selected part of the program; if nothing is selected, no action is performed. In contrast, when an action
is called via the context menu and no line is selected in the program, the action is performed for the line that you
right-clicked.
Note that any actions that modify the position of the PC will cause the call stack to pop all procedure calls until
the current procedure call remains on top. This is relevant in case the current procedure call is not the top-most
procedure call and is necessary to secure the consistency of the call stack. Modification of the PC can happen as
well directly as described above or indirectly by, for example, inserting a program line above the PC in the current
procedure body.
Procedures can be created from scratch or from selected program lines in the currently displayed procedure. When
you start a new HDevelop program, there is only the main procedure. As the program grows, you often find that
chunks of code have to be reused or they constitute a functional unit. In these cases, it is good practice to relocate
the corresponding lines to a new procedure.
As an example, consider the following example program:
In this example, you want to reuse the selected program lines. To create a new procedure from these program lines,
click the menu entry Procedures . Create New Procedure (page 59) or select the corresponding entry from
the context menu of the program window. The new procedure can now be setup in the procedure interface dialog.
GUI Reference
See also: Procedures . Create New Procedure (page 59) / Edit Procedure Interface (page 59).
Name The procedure name must start with a letter and may consist of alphanumeric characters and underscores.
If you enter an invalid name, such as an operator name, a reserved word, or a name that contains invalid
characters, the text field will be highlighted. You will not be able to close the dialog and apply the changes
until a valid name is provided.
Operator names cannot be used because operators and procedures share the same namespace. However,
procedures with the same name (but different locations) are allowed in HDevelop (see section 5.5 on page
46).
Password You can optionally protect procedures by a password. Protected procedures can be used in HDevelop
programs without restrictions. However, to view and modify them the correct password needs to be provided.
See section 5.6 on page 47 for more information.
Type This check box determines the procedure type (see section 5.1 on page 43). By default, a local procedure
is created. Local procedures are saved within the HDevelop program. External procedures are saved as
stand-alone files. Libraries may contain multiple procedures in a single file.
The file type of external procedures can be specified explicitly (.hdvp or .dvp; see section 5.2 on page 44).
External and library procedures can be reused in other HDevelop programs. You can change the procedure
type at any time.
136 Graphical User Interface
If the selected directory is not currently contained in the configured directories, HDevelop will ask you if
you want to add it to the list when you commit the dialog.
Library For library procedures a target library has to be specified. The list contains all currently available li-
braries. The buttons next to the list allow you to create a new library or browse for an existing library that is
currently not configured.
Scope (Public or Private)
Specifies whether procedures can be called by any procedure (public), or only by procedures in the same
directory or file (private) as described in section 5.3 on page 45. The icons of private procedures are decorated
with a green dot.
Activated This check box determines whether or not the selected external procedure can be resolved. This
option can also be toggled in the preferences (see section 6.16.6 on page 116).
Keep Explicitly Opened (existing external procedures only)
If checked, the selected procedure will be kept open for editing even if its path is not configured in the
GUI Reference
preferences. Procedures marked this way can always be selected from the drop-down list of the program
window.
Add Path (existing external procedures only)
Add the directory of the selected procedure to the list of procedure directories (see section 6.16.5 on page
115).
The addition can either be permanent or for the current session only. This is further explained in section 5.4
on page 45.
This part of the dialog is used for the definition of procedure parameters. HDevelop procedure interfaces have
the same structure as HALCON operator interfaces. They can contain parameters of the four categories iconic
input, iconic output, control input, and control output in this order. The procedure interface dialog contains four
separate areas where the different parameters types can be edited. Each area contains a button for appending new
parameters to the parameter list.
When creating a new procedure from selected program lines, HDevelop automatically determines suitable interface
parameters for the procedure from the usage of the variables in the selected code. The combo box Selection
Scheme determines the suggestion of the procedure parameters.
138 Graphical User Interface
First In If the first use of a variable inside the selected lines is as an input variable, it will be suggested as an
input parameter of the procedure.
Last Out If the last use of a variable inside the selected lines is as an output variable, it will be suggested as an
output parameter of the procedure.
All In All input variables inside the selected lines are suggested as input parameters in the procedure.
All Out All output variables inside the selected lines are suggested as output parameters of the procedure.
The classification of variables in the selected program lines is performed separately for iconic and control variables.
If a variable is an input as well as an output variable, it is assigned to the first category, and the corresponding
procedure parameter becomes an input parameter.
If, according to the above rules, a variable name would be suggested as an input as well as an output parameter, it
becomes an input parameter of the procedure. In addition, an output parameter with the variable name extended
by “Out” is created.
As an illustration, the following program lines are selected for a new procedure:
Then, based on the selection scheme All In All Out, the procedure body will read
To the right of the parameter list, the following buttons are provided:
Reset If you are creating a new procedure, clicking this button removes all entered parameters. If you are editing
an existing procedure, the original interface is restored, for example, any changes to the parameters are
undone.
6.17.5 Creating Procedures 139
Remove Using this button you can remove single parameters from the list. Before clicking this button, focus the
corresponding parameter by clicking its text field.
Move Up, Move Down Using these buttons you can alter the order of the parameters. Select a parameter by clicking
its text field and use the buttons to change its position.
To move a parameter from one parameter group to another parameter group, follow these steps:
• Remove the parameter from the current group.
• Click Apply to save the current state.
• Add the parameter to the desired group.
Vector Parameters
HDevelop 12.0 and later versions support vector variables. See section 8.6 on page 270 for a detailed description.
Vector variables can be used in procedure calls. The following buttons handle the conversion of parameters to
vector and back again. Please note that the dimension of a vector in a procedure call must match the dimension
specified in the procedure interface. If the procedure expects a vector of tuples, it cannot be called by passing a
vector of vectors of tuples.
Vector parameters are distinguished from other parameters by showing the contained type and the dimension next
to the parameter name, for example, object {1} for a vector of iconic objects, or tuple {2} for a vector of
vectors of tuples.
To Vector Set the type of the selected parameter to vector. If the selected variable is of type vector already, its
dimension can be increased with this button (label changes to Dim +).
To Object / To Tuple Set the type of the selected parameter to object (iconic parameters), or to tuple (control
parameters). If the selected variable is a multi-dimensional vector, its dimension can be decreased with this
button (label changes to Dim -).
Replace selected program lines If this option is enabled, the selected program lines are replaced by an
appropriate call to the newly created procedure see figure 6.52 for an illustration. Otherwise, the original
program lines are kept and no procedure call is added. In any case, the selected program lines are copied to
the body of the new procedure.
Adapt program (existing procedures only) This setting is not relevant for new procedures, but useful when
GUI Reference
modifying the interface of existing procedures. When enabled, all program lines calling the procedure in
question will be adapted according to the interface changes.
For example, if you decide that a certain parameter is no longer necessary, the corresponding expressions or
variable names will be removed from all procedure calls in the program when you close the dialog and apply
the changes. If this is an input parameter, the program will continue to run without further modifications.
If it is an output parameter, subsequent program lines relying on the value of that parameter will have to be
adjusted manually.
As another example, if a new parameter has become necessary, a variable of the same name will be added to
all procedure calls. If this is an input parameter, the corresponding variable will most likely not be initialized
at the time of the procedure call and has to be assigned to manually. If it is an output parameter, the program
will continue to run without further modifications.
Leaving this feature enabled is highly recommended to keep the program consistent.
OK Activating the button OK at the bottom of the dialog either creates a new procedure or commits the changes
made in the procedure interface, depending on whether the interface dialog was invoked to create a new
procedure or to modify an existing procedure. In the latter case not only the interface itself might be changed
but also the procedure’s program body and variable lists, as new variables might have been added or existing
variables might have been removed or renamed.
If you change the interface of an external procedure, be aware of the fact that other programs containing it
do not update the procedure calls. When loading these programs, the procedure calls are disabled. If the
changes were applied to a procedure that is called from inside a protected external procedure, that procedure
call is not even updated in the current program.
140 Graphical User Interface
Cancel This button dismisses the dialog. Any changes to the interface or the documentation of the edited proce-
dure are lost (with the exception of the editing status, see section 5.6 on page 47).
Apply Applies the changes in the dialog (just like pressing OK) without closing it.
Help Displays the documentation of the selected procedure. If the documentation is empty, the button will be
grayed out.
The newly defined procedure is now available for selection in the operator window. The variables that were used
to determine the procedure interface parameters are now being offered as input parameters for the procedure call.
Please note that a return call has been added at the end of the procedure body. If you create a procedure from
scratch, the newly created procedure body will contain only the return operator initially.
The combo box on top of the program window displays the name of the current procedure. You can search for
specific procedures, and select all available procedures from this box. The first element of the list will always be
the main procedure, followed by the local procedures of the current program, followed by the available external,
followed by all referenced procedures and library procedures. The procedure groups are sorted alphabetically.
Procedures can be protected with a password. Those procedures can still be selected from the list, but unless the
correct password has been entered, they will remain in a locked state. If the procedure is locked, a password button
is displayed instead of the procedure body. For more information about protected procedures, see section 5.6 on
page 47.
To view and modify the interface of the current procedure
GUI Reference
The remaining buttons provide access to the documentation of the current procedure. This is described in sec-
tion 6.17.8.
You can step through the individual tab cards of the dialog using the arrow buttons at the bottom of the dialog.
HDevelop supports the preparation of procedure documentation in the procedure interface dialog. Since procedures
are treated like operators, the same documentation slots are available. The procedure documentation is seamlessly
integrated into the online help system. For example, if you select a procedure in the operator window, clicking the
help button will take you to the corresponding page in the help window.
The documentation of the procedure can be entered in multiple languages. The language used for displaying the
procedure documentation in the online help depends on the language set in the preferences of HDevelop. To
edit the procedure documentation in a specific language, select the corresponding entry from the drop-down list
Language.
Group This is the top level element of the content hierarchy in the procedure online help. It can be used to apply
a vendor-specific tag to a group of procedures. The external procedures supplied with HALCON use the
group tag “MVTec Standard Procedures”. If no group is specified, the corresponding procedures are listed
6.17.8 Providing Procedure Documentation 143
under “No group assigned (local procedures)” or “No group assigned (external procedures)” depending on
their type.
Chapters The text fields next to Chapters can be used to specify chapter and section, so that your procedures
can be grouped thematically in the list at the bottom of the menus Menu Procedures and Menu Operators.
Note that the logical structure created by the chapter and section information does not correspond to the au-
tomatically created directory structure. At least for the external procedures, you can create the corresponding
directory structure afterwards, outside of HDevelop. The recognition of the procedures in HDevelop is still
GUI Reference
ensured, as all subdirectories of the external procedure paths are scanned as well. When editing already
existing external procedures, the changed procedures are stored in the paths they were originally found in.
Short Description Enter a short description. Usually, this should be a single sentence that describes the pur-
pose of the procedure. It appears in the overview sections of the online help of the procedures. Additionally,
the short description is displayed in HDevelop’s status bar when the procedure is selected from the menu.
Detailed Description Enter a detailed description of the procedure using Markdown syntax. For more infor-
mation about the syntax, see section 6.17.8.1.
Example
This section of the documentation is intended for code examples. This could be a working program or some
code fragments that illustrate the usage of the procedure.
Suggestions
The first field on this tab card allows you to associate keywords with the procedure.
You can specify suggested successors, predecessors, and alternatives to the current procedure.
Advanced
The text boxes in this tab card are for advanced usage only. We recommend searching the online reference
manual for usage examples.
Attention Notes about special observances when using the procedure. Supports Markdown syntax (see
section 6.17.8.1).
144 Graphical User Interface
Complexity Notes about intricate details about the procedure usage. Supports Markdown syntax (see
section 6.17.8.1).
Warning Usually used to indicate obsolete or deprecated procedures that are kept for backward compatibil-
ity. The warning text should indicate the recommended alternative.
A warning icon is displayed in the left column of the program window. Additionally, if the procedure
is selected, the warning text will be displayed in the operator window as a reminder.
Supports Markdown syntax (see section 6.17.8.1).
References with recommended reading about certain aspects of the procedure. Supports Markdown syn-
tax (see section 6.17.8.1).
Inline text formats Emphasize text with *italic* (italic) or make it strong with **bold** (bold).
~strikethrough~ is also possible.
Code blocks Insert a code block (that is, a block of monospaced text) by preceding the lines with four blanks.
Alternatively, use three backticks:
```
my code
```
1. first item
2. second item
* an item
* another item
* a nested item
The list of available procedures is loaded during the start of the Procedure Interface dialog and only updated when
the dialog is started again. Therefore, features like syntax highlighting will not work for other procedures if they
are created or changed while the dialog is open.
6.17.8 Providing Procedure Documentation 145
This section of the dialog provides tab cards for all parameters of the current procedure. The documentation
consists of a fine-grained specification of the parameters, and a short description. The specification fields depend
on the parameter type (iconic or control parameter), and on the selected semantics. In the following, the most
common fields of both iconic and control parameters are listed.
Please refer to the Extension Package Programmer’s Manual (Chapter 2.3) for additional information about the
documentation fields (especially, the semantic types).
Iconic Parameter Documentation
Field Meaning
Semantics Specifies the semantic type of the parameter.
Pixel Types Only available for Semantics image. Lists the accepted pixel types. The but-
tons Select All and None toggle the selection of all parameters.
Multi Channel Only available if Semantics = image.
False: Only the first channel of the image is processed,
True: Only a multi-channel image is accepted,
Optional: Images with an arbitrary number of channels are accepted.
Multi Value False: Only a single object (no object tuple) is accepted,
True: Only object tuples are accepted,
Optional: A single object as well as an object tuple is accepted.
Description Short description of the iconic parameter.
GUI Reference
The concept of protected procedures is explained in section 5.6 on page 47. The status of a procedure can be
changed via the procedure interface dialog (page 135). To manage the status of multiple procedures at once, click
Menu Procedures . Manage Procedures and select the tab card Manage Passwords (page 119).
6.17.9 Protecting a Procedure 147
Figure 6.60: Entering a password to protect a procedure. The option “Lock entire...” is only available for the main()
procedure.
When HDevelop is started again, the protected procedure is locked. When trying to edit the procedure, for example,
by selecting it from the combo box in the program window, a corresponding message is displayed in the program
window see figure 6.61). Additionally, a password button is displayed in the program window. The procedure
remains unlocked for this session, that is, until you close HDevelop or lock the procedure again manually.
GUI Reference
Changing the Status of a Protected Procedure
To change the status of a protected procedure, it must be unlocked. Use the procedure interface dialog (page 135)
to change the password or remove the password entirely.
Lock Lock the protected procedure and its body cannot be accessed in the current session without supplying the
password again.
Remove Removes the password. When the procedure is saved, it is no longer protected.
New password The password window is displayed and a new password can be assigned.
Cancel The operation is canceled without altering the status.
Warning
When working with protected procedures, be aware that the password cannot be reconstructed, so be very careful
not to forget it and not to repeat a typing error when assigning it!
The option “Unlock entire...” is only available if local procedures have been locked at once.
Note that, in some situations protected procedures behave differently from common procedures. In particular,
as they cannot be viewed and modified by unauthorized users, they also cannot be copied, printed, or exported
to any programming language (however, they can be duplicated using the menu entry Menu Procedures .
Duplicate...). Additionally, if a protected procedure contains a call to another procedure for which the in-
terface was changed, the procedure call is not adapted to the changes but is disabled for the current program.
6.17.10 Profiler
The built-in profiler analyzes the runtime behavior of HDevelop programs. You can use the profiling data to
evaluate the overall program execution, to optimize its performance, and to find bottlenecks.
The profiler measures the processing times in one of two modes:
• In the operator time mode, the profiler counts operator and procedure calls and measures the processing
times of operator calls.
The operator time is the appropriate measure if you plan to export your HDevelop program to a programming
language.
• In the execution time mode, the profiler measures the operator time and additionally the overhead of each
operator call inside HDevelop.
The execution time is the appropriate measure if you run the program inside HDevelop or HDevEngine.
Keep the following in mind when profiling your code:
• Disable GUI updates by calling the convenience procedure dev_update_off() at the beginning of your
program.
• Run the code to be profiled a couple of times and use the average profiling data to get more accurate results.
• Reduce the number of concurrent (background) processes.
• Avoid stop and wait_seconds instructions, or override these two operators; see Override Operator
Behavior (page 126).
To activate the profiler, click Execute . Activate Profiler (page 56).
When the profiler is activated, each program execution collects profiling data.
To reset the profiler data, click Execute . Reset Profiler (page 56),
To deactivate the profiler, click Execute . Deactivate Profiler (page 56).
The profiling results are displayed in the program window. In addition, a summarized view of the runtime statistics
is available.
6.17.10 Profiler 149
If the profiler is enabled, the tool bar of the program window contains some additional profiler-related buttons
To illustrate the way the profiler works, the HDevelop program solution_guide/basics/color_simple.hdev
is loaded and executed (see figure 6.63).
The options described below are available in the drop-down menu of the toolbar icon or the context menu of
the profiler area.
GUI Reference
Activate/Deactivate Display / Activate Only selected See section 6.17.10.3 on page 151.
Number of Calls Displays how many times each program line has been executed in total. This value is accu-
mulative until the profiler is reset.
Total Execution Time Displays the total processing time of each program line. This value is accumulative
until the profiler is reset.
Average Execution Time Displays the average processing time of each program line. This value differs from
the total processing time if the corresponding program line has been executed more than once, for example,
in a loop. The average processing time is more meaningful if the program is reset and run multiple times.
Minimum Execution Time Displays the minimum processing time of each program line. This value is only
meaningful if the corresponding program line has been executed more than once, for example, in a loop.
Maximum Execution Time Displays the maximum processing time of each program line. This value is only
meaningful if the corresponding program line has been executed more than once, for example, in a loop.
Last Execution Time Displays the processing time of the last execution of each program line.
The processing time is displayed in one of two modes as explained in section 6.17.10 on page 148, depending on
the setting in the context menu:
Execution Time Measures calls including the additional overhead inside HDevelop.
Time Display the processing times as absolute values. The unit of measure defaults to ms and is adjusted appro-
priately, for example, it switches to seconds once the value exceeds 1000ms. To specify the unit of measure
explicitly, switch to the runtime statistics window described in section 6.17.10.2.
Percentage Display the processing times as percentages. 100% refers to the accumulated times inside main, or
the currently displayed procedure, respectively.
This window displays the accumulated profiling data of the procedure that is currently displayed in the active
program window (or all procedures, see below).
The table rows can be sorted by clicking the corresponding headers. Two different display modes are available:
Operators In this mode, the profiling data is displayed per operator/procedure call, that is, multiple calls of the
same operator/procedure are summarized.
Note that the sum of all calls that comprise the program can be less than the measured duration of the main
program as the latter also includes the overhead of the operating system.
Program Lines In this mode, the profiling data is displayed per program line.
Save the runtime statistics as a plain text file (.txt or .csv). Each line contains a table row, and the columns
are separated by tabs.
Copy the selected entries to the clipboard. If no entries are selected, the entire table is copied.
Toggle the display status of the selected profiler lines (see section 6.17.10.3). This option is only available if
the display is set to “Program Lines”.
6.18 Quick Navigation Window 151
Usually, the runtime statistics window displays the profiler data of the procedure that is currently displayed in
the active program window. Turn this button on to show the profiling data of all procedures.
In general, deactivated profiler lines are not listed in the runtime statistics window (see section 6.17.10.3). If
this button is on, deactivated profiler lines will be displayed in light gray so that they can be activated again.
The profiler collects data for all executed program lines. In general, the collected data is also displayed in the pro-
gram window and the runtime statistics window. This behavior is not always desirable. In many cases, the runtime
statistics of only a small portion of the program code is relevant when evaluating its performance. Accordingly,
the display of profiler data in both the program window and the runtime statistics window can be restricted to a
selection of profiler lines.
Profiler lines can be selected in three different ways:
• Drag over a range of lines in the profiler area (see figure 6.65).
• Click a single line in the profiler area, and Shift click another line to select the range between those lines.
• Click the first line in the profiler area, and then Ctrl click additional lines to select a non-contiguous selection
of lines.
GUI Reference
In all cases, the selection is also highlighted in the left part of the program window until the mouse button is
released. This visualization supports you in selecting the desired program lines, which can be difficult if the
program window is very wide.
The values of the selected profiler lines are summed up and displayed in the status bar as shown in figure 6.66.
To display the profiler data of only the selected lines, click Activate Only selected in the context menu of the
profiler area. The result is shown in figure 6.67. Note that the gray bars are automatically adjusted to the selected
data.
To toggle the status of specific profiler lines, select the corresponding lines as described above and click
Activate/Deactivate Display in the context menu of the profiler area.
The runtime statistics window is updated to list only activated program lines as shown in figure 6.68 on page 153.
To remove additional lines from this window, select the corresponding lines and click the button. Deactivated
lines can be re-displayed in light gray if the button is turned on.
152 Graphical User Interface
The quick navigation window contains a listing of program lines that share a common property. Clicking an
entry focuses the corresponding program line in the active program window. If the selected procedure is already
displayed in a program window tab, the corresponding tab is activated. Otherwise, the current view switches to the
GUI Reference
selected procedure.
It is possible to activate, deactivate, copy, cut, or delete one or more of the selected program lines directly from the
quick navigation.
The “common properties” are available as tab cards:
This tab card lists all invalid program lines within the currently selected scope. All refers to all procedures,
Program refers to the main procedure and all procedures it uses, and Current Procedure refers to procedure
displayed in the active program window. Examples of invalid lines are unresolved procedure calls, operator or
procedure calls with a wrong number or type of parameters, or syntax errors.
Clicking an entry focuses the corresponding program line in the active program window so you fix the error. If the
selected procedure is already displayed in a program window tab, the corresponding tab is activated. Otherwise,
the current view switches to the selected procedure.
If you make changes to the program while the display of invalid program lines is still open, you will need to click
the button Refresh to update the entries.
Using the context menu, you can perform the following actions on the selected program lines:
• Copy Values (page 54)
154 Graphical User Interface
Figure 6.70: Managing invalid program lines in the quick navigation window.
This tab card provides a listing of program lines that match a previous “find all” operation. Click the button Open
Find/Replace Dialog to perform a “find all” operation. More info (page 72)
6.18.3 Breakpoints
This tab card lists all variables or program lines marked with a breakpoint (see figure 6.72). Information about the
breakpoints is given in the column Context: Breakpoints on variables list the variable name, while line-based
breakpoints list the program line (its line number is given in the column Line).
Clicking the entry of a line-based breakpoint focuses the corresponding program line in the active program window.
If the selected procedure is already displayed in a different program window tab, the corresponding tab is activated.
Otherwise, the current view switches to the selected procedure.
Clear the breakpoints of the selected program lines and immediately remove the corresponding entries from
the list. This action can also be triggered from the context menu, or by pressing Del .
Clear all breakpoints and remove all entries from the list, leaving you with an empty window.
Activate all breakpoints at once. This action can also be triggered from the context menu.
Deactivate all breakpoints at once. This action can also be triggered from the context menu.
Breakpoint on variable
GUI Reference
Line-based breakpoint
6.18.4 Bookmarks
This tab card lists all program lines that are currently bookmarked. Clicking an entry focuses the corresponding
program line in the active program window. If your program contains many bookmarks, this will be more transpar-
ent and convenient then repeatedly jumping through your selection of bookmarks with F11 and Shift+F11 . If the
selected procedure is already displayed in a program window tab, the corresponding tab is activated. Otherwise,
the current view switches to the selected procedure.
Clear the bookmarks of the selected program lines and immediately remove thecorresponding entries from the
list. This action can also be triggered from the context menu, or by pressing Del .
Clear all bookmarks and remove all entries from the list, leaving you with an empty window.
or XLDs that can be used in your current program by generating the appropriate program lines. You can also
perform simple interactive measurements in that you specify a known dimension and thus translate pixels to real-
world units.
As an introductory example, the image shown in figure 6.74 shall be used as the base image for a completeness
check application. To determine the location and the alignment of the chip in subsequent images, parts of its
imprint will be used as a landmark. Therefore, an ROI has to be created. It would be cumbersome to generate an
ROI based on written-down pixel coordinates. Instead, it is much easier to draw a figure right into the image and
let HDevelop generate the corresponding instructions.
Example image
Opened dialog
Rectangle 1
Rectangle 2
By default, the figure type is set to ROI (cf. in the tool bar), which is just what we want for this example.
• Click to create a rectangle.
Click and hold inside the graphics window to draw an initial rectangle.
You can modify the rectangle by dragging its sides or corners. Make the rectangle match the big label on the
chip (see figure 6.75 ).
Click the right mouse button to confirm the rectangle and it will be created in the toolbox.
• Repeat the last step to create another rectangle and make it match the lower part of the label (see figure 6.75
).
GUI Reference
ROI Opacity
You can adapt the opacity of the ROIs by dragging the slider.
You can make changes to a figure by clicking its individual constituent parts to select them. In figure 6.77, the
upper rectangle is selected and can be modified by dragging its handles. You can also edit the selected element in
the mode it was created in by clicking (right-click to confirm the changes).
The corresponding values in the tool box will be updated accordingly. You can also edit the values in the tool
box parametrically. The parameters Perimeter and Area are calculated depending on the other parameters of the
rectangle. They can not be modified directly and are therefore grayed out.
If you click the upper rectangle twice, the entire figure will be selected and can be dragged with the mouse.
In general, if the figure consists of overlapping elements, the first click will select the topmost element. Each
following click will select the element below the selected one. When the bottommost element is reached, the next
click will select the entire figure. In case of a very complex figure with many stacked elements it might be easier
to select a distinct element by clicking the corresponding data block in the tool box.
Mode Selection
The tool box supports three modes of operation. The modes can be selected from the drop-down button of the tool
bar or from the menu ROI Type:
(ROI) Select this mode if you want to generate a pixel-based region of interest. The ROI consists of geometric
elements. The faces defined by these elements are connected by set operations (see below).
For each element a masking mode can be toggled independently by clicking the corresponding icon in the
tool box:
The default masking mode selects the inside of the element, for example, the inner area of a circle.
The complement masking mode selects the outside of the element, for example, everything outside of a
circle.
(XLD) Select this mode if you want to generate a vector-based closed-contour XLD (or a line XLD). The XLD
consists of geometric elements. The contours defined be these elements are connected by set operations (see
below).
6.19 ROI Window 159
A B
ROI
XLD
(Path) Select this mode if you want to generate a vector-based path XLD. The path consists of geometric
elements. The individual elements of the path are connected with straight lines. This connection is done
automatically. Of the two endpoints of each element the one that is closest to an endpoint of another element
is connected to that endpoint.
Depending on the selected mode, different geometric elements are available in the tool box:
Tool Type ROI XLD Path
line X X X
circle X X -
circular arc - X X
ellipse X X -
elliptic arc - X X
axis aligned rectangle X X -
rotated rectangle X X -
arbitrary region X X -
GUI Reference
Set Operations
The elements of a figure are connected by set operations. In the case of ROIs, applying the set operations deter-
mines the final face of the compound region. For closed-contour XLDs, applying the set operations determines the
final contour of the compound XLD. Set operations are not meaningful and therefore not available in path mode.
The set operation of the next new element can be selected from the drop-down button of the tool bar or the menu
Operations. The set operation of an existing element can be changed by clicking the corresponding icon in the
element’s data block.
The following set operations are available (see figure 6.78 for an illustration):
• (union of A and B)
• (intersection of A and B)
• (difference of A minus B)
• (xor, A or B exclusively)
Interactive Measurements
Usually, all dimensions in the tool box are given in pixels. To convert pixel values to real-world dimensions, a
simple calibration can be performed, that is a known dimension has to be specified.
See figure 6.79 for an example. The length of the line is known, so it can be used for the calibration.
160 Graphical User Interface
This window displays information about the current execution status of the program.
The Thread View in the upper half displays information about all the threads that have been started (see sec-
tion 8.11.4 on page 287).
The Call Stack in the lower half contains a list of the names of all procedures that are currently called on
HDevelop’s internal call stack. The top-most procedure call belongs to the most recently called procedure, the
bottom-most procedure call always belongs to the main procedure. Clicking a procedure call in the dialog makes
the selected procedure call the current procedure call and thus the procedure belonging to the selected procedure
call the current procedure.
When you click a procedure call that belongs to a locked procedure (see section 5.6 on page 47), you can only see
the procedure body if you enter the correct password into the program window.
There are two kinds of variables in HALCON, corresponding to the two parameter types of HALCON: iconic
objects (images, regions, and XLDs) and control data (numbers, strings, handles). The corresponding variables
are called iconic and control variables. These variables can possess a value or can be undefined. An undefined
variable is created, for example, when loading a program or after inserting an operator with a new variable that is
not executed immediately into a program. You can access these undefined variables only by writing to them. If you
try to read such a variable, a runtime error occurs. If a variable obtains a value, the variable type is specified more
precisely. A control variable that contains, for example, an integer is of type integer. This type might change to
real or a tuple of integer after specifying new values for this variable. But it always remains a control variable.
6.21 Variable Window 161
Similarly, this is the case for iconic variables, which can contain regions, images, or XLDs. You can assign new
values to an iconic variable as often as you want to, but you cannot change its type so that it becomes a control
variable.
New variables are created in the operator dialog area during specification of operator or procedure call parameters.
Here, every sequence of characters without single quotation marks is interpreted as a variable name. If this name
did not exist before, the variable is created in the operator dialog area by pressing OK or Enter. The variable type
is specified through the type of the parameter where it was used for the first time: Variables that correspond to an
iconic object parameter create an iconic variable; variables for a control parameter create a control variable. Every
time an operator or procedure call is executed, the results are stored in variables connected to its output parameters.
This is achieved by first deleting the contents of the variable and then assigning the new value to it.
The variable window is similar to a watch window used in window-oriented debuggers. Inside this window you are
able to keep track of variable values. Corresponding to the two variable types, there are two areas in the variable
window. One for iconic data (above or left) and the other for control data (below or right). You can toggle the
orientation of the two parts of the variable window. To do this, double-click the dividing line between both i
parts. You can also drag that line to resize the parts.
All computed variables are displayed showing their iconic or control values (unless the automatic update has been
switched off, see section 6.16.14 on page 125). In case of a tuple result that is too long, the tuple presentation
GUI Reference
Figure 6.81: Variable window with instantiated iconic and control variables.
162 Graphical User Interface
is shortened, indicated by three dots. In this case, the full content of a variable can be displayed in an inspection
window by double-clicking the value list. See also the following sections.
In addition to line-based breakpoints (see section 6.17 on page 127), breakpoints can also be set on variables.
Right-click a variable name and select Set/Clear Breakpoint on Variable. Variables with breakpoints are
marked with next to their name.
Once a breakpoint has been set on a variable, the program execution stops whenever its value changes. A change
of value is detected in the same way the HDevelop operation != works (see section 8.5.10 on page 265):
• Iconic objects: a value change is triggered if the variable refers to a different object in the database (see also
test_equal_obj).
• Tuples: A change of the data type (for example, integer to real) does not necessarily trigger a value change.
In the following example the second assignment would not trigger a variable breakpoint on tuple:
tuple := [1,2,3]
tuple[1] := 2.0 // 2 == 2.0
Using clear_obj does not trigger a variable breakpoint. Use dev_clear_obj instead.
When a breakpoint on a variable triggers, the program execution stops and HDevelop highlights the corresponding
variable in the variable window. If the variable is not currently displayed in the selected tab of the variable window,
HDevelop switches to the tab Auto (see below).
Information about the event is displayed in the status bar, for example, HDevelop displays both the old and the
new value of a tuple.
In large programs the variable window can become quite cluttered, which makes watching selected variables
difficult. Therefore, you can customize the selection of displayed variables.
All All variables used by the current procedure are displayed at once. Global variables (see section 8.3.2) are
marked with .
Auto The variables of the current and the previous operator call are displayed. This is useful when single-stepping
through the program, because only the variables relevant to the current context are displayed.
User A user-defined selection of variables is displayed. If the tab User is active, variables can be added from a
list in the context menu. In the other tabs variables are added by selecting them first and clicking Add to
User Tab in the context menu.
It is also possible to select a variable name in the program window and drag it to the variable window. The
corresponding variable will then be added to the watched variables, and the tab User will be activated.
Global All global variables (see section 8.3.2) of the current program are displayed. This includes global variables
in external procedures, even if they are not used in the current program.
The context menu of the global variables includes an additional entry Declared in. It lists the names and
line numbers of the procedures that declare (and thus use) the corresponding global variable. Click an entry
to jump to the corresponding location in the program window.
6.21.3 Iconic Variables 163
Figure 6.82: Displaying information about an iconic variable with a reduced domain.
The iconic variables are represented by icons, which contain an image, a region or an XLD, depending on the
current value. The icons are created depending on the type of data according to the following rules:
• For images the icon contains a zoomed version of them, filling the icon completely. Due to the zooming onto
the square shape of the icon, the aspect ratio of the small image might be wrong. If there is more than one
image in the variable, only the first image is used for the icon. Similarly, for multi-channel images only the
first channel is displayed. An exception is made for images with three channels: These are displayed as color
icons (RGB).
The domain of the image is not reflected in the displayed icon. Information about the domain can be obtained
from the tool tip which appears when the mouse cursor points to the icon. See figure 6.82 for an illustration.
• Regions are displayed by first calculating the smallest surrounding rectangle and then zooming it so that
it fills the icon using a border of one pixel. In contrast to images, the aspect ratio is always correct. This
GUI Reference
can lead to black bars at the borders. The color used to draw the region is always white without further
modifications (except zooming).
• XLD data is displayed using the coordinate system of the largest image used so far. The color used for XLD
objects is white on black background.
Because of the different ways of displaying objects, you have to be aware that the coordinates cannot be compared.
The variable name is positioned below each icon. They are displayed in the variable window in the order of
occurrence or name from left to right. If there is not enough space, a scrollbar is created, which you can use to
scroll the icons.
• Images: The image type and size and the number of channels is displayed. If the iconic variable contains
multiple images, the properties of the first image are reported.
• Regions: The area and the center of the region is displayed. If the iconic variable contains multiple regions,
the properties of the first region are reported.
• XLDs: The number of contour points and the length is displayed. If the iconic variable contains multiple
XLDs, the properties of the first XLD are reported.
Context Menu
Clicking an icon with the right mouse button opens a context menu with several options:
Display: Display the selected iconic variable in the active graphics window. Regions and XLDs are displayed on
top of the previous contents of the active graphics window.
Display Channel: Display a single channel of the selected iconic variable in the active graphics window. This
menu lists up to 15 channels. If the iconic variable contains more than 15 channels, you can access the
remaining channels by clicking “Select...”, which opens an auxiliary window listing all channels.
This entry is only available if the selected iconic variable contains a mult-channel image.
Display Content: Display a single object of the selected iconic variable in the active graphics window. This
menu lists up to 15 objects. If the iconic variable contains more than 15 objects, you can access the remaining
objects by clicking “Select...”, which opens an auxiliary window listing all objects.
This entry is only available if the selected iconic variable contains multiple objects, for example, multiple
images, regions, or XLDs.
Clear / Display: Clear the active graphics window before displaying the selected iconic variable.
Clear Variable: Clear the selected iconic variable. The contents of the variable become undefined.
Save: Save the contents of the selected iconic variable to a file. Depending on the content type of the vari-
able (image, region, XLD, ...), different file formats are offered. See also: write_image, write_region,
write_contour_xld_dxf, write_polygon_xld_dxf, fwrite_serialized_item.
Set/Clear Breakpoint on Variable: Toggle a breakpoint on the variable under the mouse cursor (see sec-
tion 6.21.1 on page 162).
Activate/Deactivate Breakpoint on Variable: Toggle activation of breakpoint on the variable under the
mouse cursor.
Add to User Tab: The selected variable is added to the tab User.
Find Variable: Open the Find/Replace... (page 72) dialog with the name of the selected variable prese-
lected.
Insert dev_display() into program: Insert the operator dev_display into the program window at the IC.
The parameter of the inserted instruction is the name of the selected iconic variable.
Shortcut: Ctrl + double-click variable icon.
Declared in (global variables only) List the names and line numbers of the procedures that declare (and thus
use) the selected global variable. Click an entry to jump to the corresponding location in the program
window.
Channel 1: red
Channel 2: green
Channel 3: blue
Update Variables: Toggle whether variables will be updated during program execution. This is the same setting
as in the runtime preferences (see section 6.16.14 on page 125).
GUI Reference
Add Variable (tab User only): This submenu contains a list of all variables that are currently not displayed in
the tab User. Click a variable name to add the corresponding variable to the tab.
Remove from User Tab (tab User only): The selected variables are removed from the tab User.
You can display the corresponding iconic variable in the active graphics window (with or without clearing the
window first), and you can clear the iconic variable. If an iconic variable contains multiple items, you can also
select a specific item from a submenu. Up to 15 items are listed in this menu. If an iconic variable contains
more than 15 items, the remaining items can be accessed by clicking Select.... If you click Select... in this
submenu, you can quickly browse the items of the iconic variable from a dialog. This also works for multi-channel
images. See figure 6.83 for an example.
• A last exception is an empty region. This is one region that does not contain any pixels (points), that is the
area (number of points) is 0. You must not confuse this case with the empty tuple, because there the area is
not defined. The empty region is symbolized by an empty set icon .
To the right of the variable name you find its values in the default representation. If you specify more than one
value for one variable (tuple), they are separated by commas and enclosed by brackets. If the number of values
exceeds an upper limit, the output is clipped. This is indicated by three dots at the end of the tuple.
Context Menu
Inspect: Inspect the values of the selected control variables in an auxiliary window (see section 6.22). The
display depends on the semantic type of the selected control variable:
• Tuples
• Vectors (page 169)
• Handles (page 170)
• Matrices (page 170)
• Poses (page 170)
• Image acquisition devices (page 171)
• Functions (page 171)
• 3D object models (page 176)
Inspect as Handle Force the generic handle inspect window for all types (see section 6.22.3 on page 170).
Inspect as Tuple Force the tuple inspect window for all types (see section 6.22.1).
Plot as Function Plot the tuple values (see section 6.22.7 on page 171). The variable is assumed to contain y
values for equidistant x values. This entry is only available for numeric tuples.
Plot as X/Y pair(s) Generate a scatter plot of two variables selected using Ctrl . This entry is only available
if two variables containing numeric tuples of equal length are selected.
Copy: Copy the values of the selected variables to the system clipboard. If the variable window has the keyboard
focus, Ctrl-C can be used as a shortcut. Tuples with zero or more than one values are returned in tuple
notation: [.., ..]. If multiple variables are selected, the tuples of the different variables are separated by
a new line.
Clear Variable: Clear the selected control variables. The contents of the variables become undefined.
Save: Save the contents of the selected control variable to a file. Depending on the content type of the variable
(tuple, handle, ...), different file formats are offered. See also: write_tuple, write_bar_code_model,
write_calib_data, write_data_code_2d_model, write_matrix, write_object_model_3d.
Set/Clear Breakpoint on Variable: Toggle a breakpoint on the variable under the mouse cursor (see sec-
tion 6.21.1 on page 162).
Activate/Deactivate Breakpoint on Variable: Toggle activation of breakpoint on the variable under the
mouse cursor.
Add to User Tab: The selected variables are added to the tab User.
Find Variable: Open the Find/Replace... (page 72) dialog with the name of the selected variable prese-
lected.
Declared in (global variables only) List the names and line numbers of the procedures that declare (and thus
use) the selected global variable. Click an entry to jump to the corresponding location in the program
window.
6.22 Variable Inspect 167
Control variables containing tuples are displayed in a tabular format (see figure 6.84). This is especially useful for
the inspection of tuples with a large number of values.
In addition to the tuple values, some statistical data can be displayed:
GUI Reference
Variable names
• minimum value
• maximum value
• sum of values
• mean value
• deviation
• value types
• number of values
• semantics (if appropriate).
You can select, which statistical data is displayed by right-clicking the statistics table and selecting the correspond-
ing entries.
The selected
values of an inspection window can be copied to the system clipboard using the context menu or
pressing Ctrl+C . The columns of the copied values are separated by tabs and the rows are separated by newlines,
or presented as an HTML table (depending on the application the values are pasted into).
Variable Inspect
1
2 -> [2,22,222]
3
1
2
22
222
3
Do not forget to include the original value (2) if you want to extend the tuple.
Additional values can be appended at the end of the tuple by entering them into the column labeled +.
To delete a variable value, press Delete or enter an empty tuple:
Variable Inspect
1
2 -> []
3
6.22.2 Inspecting Vectors 169
1
3
Vectors are container variables for tuples, iconic objects, or vectors. See section 8.6 on page 270 for a detailed
description. Because vector variables can be multi-dimensional, the inspection window presents their values in a
tree-like fashion.
The example variable vector is a two-dimensional vector. It contains three vector variables which in turn contain
GUI Reference
two single-element tuples each. Figure 6.85 shows the inspection window of this variable. Depending on the
dimensionality of the inspected vector variable, the sub-nodes can be further inspected by clicking the icon in
front of the corresponding element. If the inspected vector variable contains control values as in the example, the
individual values at the leafs of the tree can be edited like tuple variables (see section 6.22.1 on page 168).
In addition to the tuple values, some statistical data can be displayed:
• minimum value
• maximum value
• sum of values
• mean value
• deviation
• value types
• number of values
• semantics (if appropriate)
• vector dimension.
You can select the statistical data to be displayed by right-clicking the statistics table and selecting the correspond-
ing entries.
170 Graphical User Interface
Handles are references to complex data structures. The handle inspect window displays the attributes and values
of the selected handle variable. The values are read-only, they cannot be modified in the dialog. They are updated
automatically if they change in the background.
Depending on the semantic type of the referenced data, specific inspection windows are available which are de-
scribed in the following sections. Double-clicking a handle variable with no specific inspection window always
opens the handle inspect window.
When inspecting handles that contain 3D object models, those can be visualized directly from within the handle
inspect window by checking the checkbox next to them. This is also possible for 3D object models that are
contained in, for example, dictionaries. Note that when visualizing multiple 3D object models from within a
handle inspect window, all those models will be shown in a single visualization window.
Control variables that reference a matrix are displayed in a tabular format as displayed in figure 6.87.
The selected
values of an inspection window can be copied to the system clipboard using the context menu or
pressing Ctrl+C . The columns of the copied values are separated by tabs and the rows are separated by newlines,
or presented as an HTML table (depending on the application the values are pasted into).
Control variables that reference a pose are displayed in a tabular format as displayed in figure 6.88.
The selected
values of an inspection window can be copied to the system clipboard using the context menu or
pressing Ctrl+C . The columns of the copied values are separated by tabs and the rows are separated by newlines,
or presented as an HTML table (depending on the application the values are pasted into).
6.22.6 Inspecting Image Acquisition Device Handles 171
For an image acquisition device handle, a dialog with basic image acquisition device parameters is opened (see
figure 6.89). It displays the name, image size, device, port, and other features of the image acquisition device.
GUI Reference
The toggle button Online allows you to grab images continuously and to display them in the active graphics
window. Multiple online inspections from different image acquisition devices at the same time are also supported
by opening additional graphics windows before clicking the corresponding button Online. If an error occurs
during grabbing, it is displayed in the status bar of the dialog.
Control variables with the semantic type function_1d are displayed as a function plot by default upon inspection.
It is also possible to plot the data of variables containing arbitrary numeric tuples (length > 1) by selecting Plot
as Function from the context menu of the variable window. It is assumed that these variables contain y values
for equidistant x values. Finally, you can select two variables containing numeric tuples of equal length to generate
a scatter plot (see below).
Example:
The variables X and SinX store discrete input values for the sine function in the range [−π, π].These values are fed
into the operator create_funct_1d_pairs to create the function variable F with the semantic type function_1d.
172 Graphical User Interface
Plot area
Plot list
x-axis
y-axis
Double-click F in the variable window to generate a function plot (see figure 6.90).
Right-click SinX in the variable window and select Plot as Function to generate a function plot of the y values
(see figure 6.91). Note that the tuple index (0 .. 36) is used to set the x range. The plot style of numeric tuples
defaults to steps. This can be changed from the context menu of the plot list (see below).
Action Shortcut
Description
Reset Bounds w or Ctrl+Shift+W Set width to default.
6.22.7 Inspecting Functions 173
Action Shortcut
Description
Reset Bounds h or Ctrl+Shift+H Set height to default.
Set 1:1 Aspect Set y range to same scale as x range.
Ratio
Linear Scale Set y-axis to linear scale.
Logarithmic Set y-axis to logarithmic scale.
Scale
User-defined Freeze y range.
Range
Increasing Range Let y range grow on demand.
Adaptive Range Let y range grow or shrink on demand.
GUI Reference
174 Graphical User Interface
Action Shortcut
Description
Reset Bounds r or Ctrl+Shift+R Reset width and height to default.
Zoom In Mode + or Ctrl+Shift++ Zoom in (both axes).
Zoom Out Mode - or Ctrl+Shift+- Zoom out (both axes).
Enter Bounds... b or Ctrl+Shift+B Enter vertical and horizontal bounds para-
metrically.
Show Mouse p or Ctrl+Shift+P Visualize the mouse position with a cross
Position hair.
Show Function x or Ctrl+Shift+X Display the function value at horizontal
Value At X mouse position.
Show Function y or Ctrl+Shift+Y Display the function value at vertical
Value At Y mouse position.
Show Background g or Ctrl+Shift+G Display grid lines in the background of the
Grid plot area.
Insert Plot Code Ctrl+Shift+V Generate code to plot the current view in a
for Graphics graphics window.
Window
Scatter Plot
A scatter plot plots values of one variable against values of another. Both variables must have the same length.
There are two ways to generate a scatter plot:
• Select two variables containing numeric tuples of equal length in the variable window using Ctrl . Select
Plot as X/Y pair(s) from the context menu. The first variable in the current sorting order is used as the
x variable.
• Select the first variable in the variable window. Select Plot as Function from the context menu. In the
plot list of the function plot add the other variable. Open the context menu of the variable containing the y
values, select Plot against X values, and select the other variable.
The context menu of the plot list contains additional entries for the variables used for a scatter plot:
As an illustration we add the following assignment to the above code lines, and select the variables SinX (sine
function) and CosX (cosine function) to generate the scatter plot from figure 6.92.
6.22.7 Inspecting Functions 175
Figure 6.92: Example for a scatter plot using the variables SinX and CosX.
Bars Steps
GUI Reference
Lines Dots Lines and Dots
CosX := cos(X)
Note that the aspect ratio of the plot is chosen to match the window size. To get a correct aspect ratio (in this case:
a round plot), select Set 1:1 Aspect Ratio from the context menu of either the x- or the y-axis.
Plot Style
The available plot styles are illustrated in figure 6.93.
The first two styles (bars and steps) are only available for plots of numeric tuples against their index. The other
styles (lines, dots, lines and dots) are available for all function plots.
176 Graphical User Interface
After executing this line, double-clicking the control variable ObjectModel3D opens the inspection window from
figure 6.94. See get_object_model_3d_params for a description of the displayed properties.
To visualize a specific 3D object model or all 3D object models (in case the handle contains multiple models),
select the checkbox of the corresponding model ID or the checkbox next to Display Models, respectively. The
selected models are then visualized in a special visualization window shown in figure 6.95.
The slots Color and Color Normals of the parametric inspection window define the color of the displayed
model(s) or the optionally displayed normals, respectively. These colors can be changed by double-clicking the
corresponding slot and selecting a color name from the drop-down list. You can also enter a valid color name (page
321) or an arbitrary hex triplet into the color slots.
The 3D object model visualization window supports the following interactions:
drag inside circle rotate around X/Y axis
drag outside circle rotate around Z axis
mouse
wheel zoom in and out
Shift +drag
zoom in and out
Ctrl +drag pan 3D object model(s)
Attributes: Select the display mode of the 3D object model(s). Possible values are: auto (chooses the most
appropriate mode), faces, primitive, points, and line.
Alpha: Set the translucency of the displayed 3D object model(s). The alpha value ranges from 0 (full trans-
parency) to 1 (opaque).
The tool bar buttons control the way the 3D object models are displayed:
GUI Reference
See also: Menu Visualization . Zoom Window.
The zoom window enables the interactive inspection of image details. You can open up any number of zoom
windows with different zoom levels (see Menu Visualization . New Zoom Window). The window also displays
the gray values of each image channel at the mouse cursor position. Apart from this, the pixel type, the number of
channels, and the current position of the mouse cursor are displayed. The percental scale can be selected from the
combo box. It is related to the original size of the image.
There are multiple methods to navigate the zoom window:
Check Follow Mouse and move the mouse pointer over the image to select the zoomed area. Click once to keep
the currently displayed area in the zoom window, when the mouse cursor moves out of the image window. Or,
uncheck Follow Mouse and click (or drag) inside the image to select the zoomed area. The red square in the
center of the zoom window indicates the position of the mouse cursor. The corresponding coordinates are also
displayed at the bottom of the window.
You can select a particular pixel by clicking it. The zooming tool stores this position internally, and will redisplay
the thus selected part of the image object when you leave the graphics window. This enables you to have a
meaningful display in the zooming tool whenever you want to do actions outside of the graphics window.
For finer control of the zoomed area,click inside the zoom window to give it the focus and use the cursor keys to
move pixel-wise. Press and hold the Alt key and use the cursor keys to move ten pixels at a time. Click inside the
zoom window to move relative to the center position. For example, clicking ten pixels above the center will move
the view up ten pixels.
178 Graphical User Interface
The lower part of the window contains a gauge to display the gray value of the center pixel graphically. The range
goes from 0 (left) to 255 (right). Normally, the gray value of the first channel is displayed with a black bar. For
images with multiple channels the gauge is split accordingly to show individual bars for each channel. Thus, for
color images in RGB-space (three channels with red, green, and blue values) three colored bars are used. If the
gray value is below 1, the gauge is light gray (background). If the value is above 255, the gauge is dark gray or
colored for RGB images.
Above the gauge, the gray values are displayed as numbers. Up to five channels are displayed. If more than five
channels are present, the remaining channel values are truncated.
Next to the gauge, the coordinates of the mouse position are displayed. Below these, the image size, pixel type,
and the number of channels are shown.
The button next to the scale combo box enlarges the zoom window so that partially visible pixels at the border
become fully visible.
HDevelop Assistants 179
Assistants
Chapter 7
HDevelop Assistants
HDevelop contains assistants for specific machine vision tasks. Each assistant provides a user interface tailored to
the requirements of its task. Using this interface, you can interactively set up and configure the assistant to solve a
specific machine vision problem. When the configuration is working satisfactorily, the assistant can be instructed
to generate HDevelop code into the current program. You can also save an assistant’s configuration for later use.
File . Load Assistant Settings... Using this entry, a previous configuration can be loaded from a file
which has been generated using the menu entry Save Current Assistant Settings....
File . Save Current Assistant Settings... You can save the configuration of an assistant to a file for
later use. The default extension for these configuration files is .das.
File . Close Dialog The assistant is closed, but the current configuration is preserved. This menu entry per-
forms the same function as the assistant’s close button. You can restore a closed assistant by clicking the
numbered entry in the menu Assistants which is generated when a new assistant is opened.
File . Exit Assistant The assistant is quit. The resources used by the assistant are released. The link to the
generated code is lost. It is not possible to restore the assistant unless the setup has been saved to a file. The
menu entry in the menu Assistants is also removed.
Code Generation . Insert Code for Selection Insert HDevelop code based on the current settings of the
assistant. The code is inserted at the IC. As long as the associated assistant is not quit, you can change the
settings and update the code accordingly.
Code Generation . Release Generated Code Lines The link to the generated code is cut off. The code
remains in the program, but can no longer be updated or removed from the (formerly) associated assistant.
Nevertheless, you can generate new code with the current settings of the assistant.
Code Generation . Delete Generated Code Lines The generated code is deleted from the program. Please
note that any manual changes to the generated lines are deleted as well.
Code Generation . Show Code Preview Generate a preview of the code based on the current setup of the
assistant. If the program already contains generated code which is linked to the current assistant, the changed
code lines can be compared side-by-side in the preview.
180 HDevelop Assistants
Image File(s)
Activate this radio button to load images from files. You can enter the names of image files in the text field.
Multiple file names are separated by a semicolon “;”. If an image with no path name or a relative path name
is given, the image files are searched in the directories specified by the environment variables HALCONROOT and
HALCONIMAGES.
You can also enter the full path of an image directory to specify all images of the given directory. If the check box
Recursive is ticked, the images of all subdirectories are specified as well.
Pressing Return will display the first of the specified images in the active graphics window.
The buttons Select File(s) ... and Select Directory ... open a file browser to select multiple images
or an image directory, respectively. After clicking OK in the file browser, the text field is updated with the selected
items, and the first image is displayed in the active graphics window.
Use the entry Snap or Live in the menu Acquisition, or the corresponding tool bar buttons to view the selected
images one after another.
This tab card is only available if the image source is set to an image acquisition interface. The connection pa-
rameters are described below. See the description of the operator open_framegrabber for additional information
about the fields.
Configuration
Device Select the ID of a board, camera, or logical device if multiple devices are available for the selected image
acquisition interface.
Upon building the list of devices, the assistant queries the status of each device. Depending on the image
acquisition interface, devices may be reported as misconfigured. If you select such a device, the assistant may
suggest a Generic parameter that potentially resolves the misconfiguration. If you confirm this suggestion,
the parameter will be entered into the Generic slot (see below). If a device is labeled with a question mark
icon, it is either read-only, busy, or unknown.
Port Specify the ID of the input port.
7.1.3 Tab Parameters 181
Assistants
Trigger Tick the check box if the image acquisition is controlled by an external trigger.
Resolution (X / Y) Specify the factor for image width / height.
Color Space Specify the configuration for color acquisition.
Field Specify the frame selection for interlaced cameras.
Bit Depth Specify the number of bits used for one image channel.
Generic Some image acquisition interfaces support device-specific parameters to preset selected values before
the camera is initialized. The parameters the interface claims to support are suggested as a drop-down list.
To set a generic parameter, select it from the list, and edit the assigned value, that is, the value after the =.
Multiple generic parameters may be set by separating the entries with a comma.
If the selected image acquisition interface does not support generic parameters, this field is grayed out.
See the documentation of the individual image acquisition interfaces for more information about the sup-
ported generic parameters.
Action Buttons
When the connection parameters are set up, the action buttons are used to connect to and acquire images from
the specified device. Messages about connection errors are displayed in the status line of the image acquisition
assistant window.
Connect Connect to the specified image acquisition device. If the connection fails, carefully check the configura-
tion in the above fields. Not all combinations of settings are allowed for all devices. It is recommended that
you enable low level error messages (see General Options . Experienced User) to find out what is go-
ing wrong. Please note that an established connection is closed automatically, if the connection parameters
are modified.
When the connection is established, this button can be used to disconnect the device.
Snap Acquire a single image from the device (first connecting to the device if needed). The image is displayed in
the active graphics window unless Display Image is set to Disabled (see Inspect).
Live Start/stop live image acquisition mode. The images are displayed in the active graphics window unless
Display Image is set to Disabled (see Inspect).
Detect Clicking this button will attempt to redetect valid parameters for the current device.
Reset All Reset all connection parameters to their default values.
Interface Library The image acquisition interface library (DLL or shared object) used by the current connec-
tion is displayed in this field.
Update Image If this check box is ticked, a new image is acquired immediately after each parameter change.
Disable the check box if you want to change multiple parameters at once.
Refresh Refreshes the list of supported parameters and their value ranges. This is useful for parameters with side
affects.
Reset All Resets all parameters to their default values. Individual parameters can be reset by clicking the corre-
sponding button displayed to the right of each parameter.
182 HDevelop Assistants
Parameter Grouping
The available parameters are grouped by user parameters, read-only parameters, action parameters and write-only
parameters. The latter cannot be changed in the assistant and are listed only for reference. The parameters of
some of the interfaces are additionally grouped by category and visibility. If grouping information is available, the
amount of displayed parameters can be reduced by choosing a subject matter from the down-down list Category.
You can further filter the parameters by selecting a skill level from the down-down list Visibility (beginner,
expert, or guru).
By default the parameters are sorted thematically. You can also sort the parameters by name (check box Sort by
Name).
Setting Parameters
The parameters are displayed in a tabular format. The first column shows the parameter name. The second column
depends on the parameter type:
• If the parameter is editable, its value can be entered into a text field. This field may contain value suggestions
as a drop-down list. Numeric values can be incremented/decremented using the arrows next to the text field.
• If the parameter is read-only, its value is displayed, but cannot be modified.
• For action parameters, the corresponding action can be triggered by clicking the Apply button.
The third column is reserved for numeric parameters. It contains a slider to quickly alter the parameter value within
the defined range. Please note that low level error messages are suppressed while dragging the slider. If no range
is defined, no slider is displayed.
The fourth column contains a reset button for editable parameters. Click it to reset altered parameters to their
default value.
Display
Display Image It is recommended to set the display mode to Normal unless you wish to make speed measure-
ments. Other modes are Volatile (volatile grabbing), and Disabled (grabbing images without displaying
them).
Output Window Specifies the graphics window for the image display (either the active graphics windows, or a
window ID from the list).
Statistics
This area displays statistics for the acquisition time and the frame rate of all acquired images.
Ignore first image of live acquisition When the first image is acquired in live mode, a certain amount
of overhead is added. Therefore, it is recommended to check box to increase the accuracy of the results.
Status Bar
Show frames per second during live acquisition Usually, the number of grabbed images and the ac-
quisition time of the last image are displayed in the status bar of the window. Ticking this check box causes
the frame rate (frames per second) to be displayed in live mode.
Show Min/Mean/Max If selected, the image acquisition statistics are also displayed in the status bar of the win-
dow. This way, you can watch the acquisition statistics in any tab card of the assistant.
The settings made in the tab cards Source, Connection, and Parameters can be distilled to program lines
that perform the desired image acquisition in your current program. The fields in this tab card specify the code
generation details. You can preview the code lines in the panel Code Preview. This panel can be toggled between
hidden and displayed state by clicking the button next to the panel label.
7.1.6 Menu Bar 183
Acquisition
The settings of this section are available if images are acquired from an image acquisition interface.
Assistants
Control Flow: The initialization code for the selected image acquisition interface is always generated (setting
Initialize Only). It opens a connection to the specified image acquisition device, and sets all changed parame-
ters. You can also generate code to acquire a single image (setting Acquire Single Image), or to acquire images
in a loop (setting Acquire Images in Loop).
Acquisition Mode: You can switch between synchronous and asynchronous acquisition. The latter runs in the
background and is recommended for continuous acquisition.
Variable Names
This section defines the variable names that are used in the generated code.
Connection Handle: Variable storing the acquisition handle. The image acquisition interface is accessed by this
name. Set to AcqHandle in the example below.
Image Object: Variable used for the acquired images. Set to Image in the example below.
The following variables have to be specified if Source is set to Image File(s) and multiple files are specified:
Loop Counter: Variable storing the loop index.
Image Files: Variable for storing the image names as a tuple.
Example Code
* Code generated by Image Acquisition 01
open_framegrabber ('GigEVision', 1, 1, 0, 0, 0, 0, 'progressive', 8, 'gray', \
-1, 'false', 'default', '003053095003_Basler_scA160014gc', \
0, -1, AcqHandle)
grab_image_start (AcqHandle, -1)
while (true)
grab_image_async (Image, AcqHandle, -1)
* Do something
endwhile
close_framegrabber (AcqHandle)
Menu Acquisition
Connect Connect / disconnect the selected image acquisition device. See section 7.1.2 on page 180.
Most applications that need a previous calibration of the camera system belong to the area of 3D machine vision.
These applications require a 3D model of the camera system. Calibration is necessary to gain information about
distortions (perspective and lens distortions) in an image and about parameters of the camera system. Calibrating
your camera system with the HALCON Calibration Assistant enables you to measure in the world coordinate
system with a high accuracy. This task can be performed by taking images of a known object, a calibration plate.
The Calibration Assistant of HDevelop is a front-end to HALCON’s operator camera_calibration. Using the
Calibration Assistant you can
All you need is a set of suitable calibration images (the number and requirements depend on the used calibration
plate, please see the reference manual chapter “Calibration”). The Calibration Assistant then returns the calibration
results and enables you to generate code and insert it into a given program.
The Calibration Assistant can calibrate vision systems based on standard lenses as well as on telecentric lenses.
With the HALCON Calibration Assistant you can
• perform a calibration,
• view the calibration results (page 195),
• generate code (page 196) for the calibration or for using the calibration results and insert it into a program
for further use in a subsequent application.
A reference to the elements of the Calibration Assistant can be found in the Calibration Assistant Reference (page
198).
For further information about camera calibration, please refer to the reference manual chapter “Calibration” or the
corresponding chapter in the solution guide on 3D Vision.
ATTENTION: Keep your camera setup (aperture, focus, pose) fixed when you have chosen it! This applies
to the calibration process itself as well as to the subsequent application. Any changes will result in the failure
of the calibration or, even worse, in wrong output values.
In this guide, the following special terms are used:
Calibration By calibrating (page 188) a vision system, you extract information about it, for example, its focal
length or its position and orientation relative to the “world”. However, even with such information you
cannot fully reconstruct the 3D world from a single image. For example, you can determine the (3D) size
of an object only if you know its distance from the vision system (when using a standard lens). Calibration
is a preparation for all subsequent image processing applications. The Calibration Assistant needs to grab a
set of images of a special calibration object placed in front of your vision system. You can choose between
a Full Calibration and a User-Defined Calibration, where known parameters are not calibrated again.
Calibration Plate This is an object whose shape is known precisely. Two different types of standard HALCON
calibration plates are available: Calibration plates with hexagonally arranged marks and calibration plates
with rectangularly arranged marks. Transparent calibration plates are available for applications requiring
backlight illumination. Additionally, the calibration plates are available in different sizes. Which calibration
plate is suited best depends on your machine vision task: As a rule of thumb, if you grab an image of the plane
of measurement, calibration plates with hexagonally arranged marks should fill the whole image and calibra-
tion plates with rectangularly arranged marks should fill a fourth of the image. The bigger calibration plates
(160mm and 320 mm for calibration plates with hexagonally arranged marks and 100mm and 200mm for
calibration plates with rectangularly arranged marks, made from aluminum) come together with a file con-
taining their exact measurements (calplate_160mm.cpd, calplate_320mm.cpd, caltab_100mm.descr,
and caltab_200mm.descr). Please copy this file to the subdirectory calib of the HALCON base directory
7.2.2 How to Calibrate With the Calibration Assistant 185
you chose during the installation. This is not necessary for smaller (ceramics) calibration plates as they can
be manufactured very precisely and can therefore use standard description files (.cpd files for calibration
Assistants
plates with hexagonally arranged marks and .descr files for calibration plates with rectangularly arranged
marks). If you use your own calibration plate, you have to create the description file yourself and copy it
into the subdirectory calib.
Calibration Plate Extraction Parameters These parameters (page 194) influence the extraction of the calibra-
tion plate. You may change them to improve the extraction of the plate if necessary. We recommend,
however, that you try to improve your image quality first.
Camera Parameters Internal Camera Parameters describe the camera itself, for example, its Focal Length,
Cell Width and Cell Height. These parameters are part of the calibration results, initial values for some
of them are also needed for the setup of the calibration.
Camera Pose The position and orientation of the world coordinate system relative to the camera are called the
external Camera Parameters. They are part of the calibration results.
Display Parameters On the Calibration (page 188) tab, you can choose the display parameters, like colors, as
you prefer them. See also Display Parameters (page 194).
Full Calibration In a Full Calibration, the complete camera system is calibrated. The only information
needed are approximate values for Camera Type, Cell Width, Cell Height and Focal Length as well
as the question whether you are using a Telecentric camera (in which case the Focal Length is not
required).
Image Rectification Based on the calibration results, you can remove image distortions. This is called image
rectification. Example code is available from the Code Generation tab (page 196).
Pose Estimation When the interior parameters are calibrated, it is possible to estimate the camera pose from a
single image. Example code is available from the Code Generation tab (page 196).
Reference Image This image locates the world coordinate system, which then has its origin at the origin of the
calibration plate in the reference image. The origin of the calibration plate is the center of the central
mark of the first finder pattern for calibration plates with hexagonally arranged marks and the middle of the
calibration plate for calibration plates with rectangularly arranged marks. By default, the first calibration
image is used as the reference image. However, you can choose any other image of the calibration sequence.
Standard Lenses A standard lens is similar to the one in the human eye: It performs a perspective projection;
hence, objects become smaller in the image the further they are away.
Telecentric Lenses Telecentric lenses perform a parallel projection. Therefore, objects have the same size in the
image independent of their distance to a camera. This means that they can lie in different planes; only the
orientation of the planes relative to the camera must be identical.
User-Defined Calibration The setup step Calibration Task provides a User-Defined Calibration, which
enables you to perform calibrations with special setups or re-use parameters from previous calibrations.
World Coordinates Measurements and XLD contours can, after finishing the calibration, be transformed into
(3D) world coordinates, meaning the coordinates of the world (for example, in millimeters), as opposed to
those of an image (in pixels). Example code is available from the Code Generation tab under Sample
Usage (page 197).
Quality Issues A high quality of the calibration images is essential not only for the calibration itself but for the
quality of the calibration results. Examples for bad image quality are overexposure of the calibration plate,
bad mark contrast or very small mark size. These quality issues are listed under Quality Issues (page
191) on the Calibration tab. Sorting out images with too many defects improves the calibration results.
For the calibration setup in the Setup tab, the basic information has to be filled in. Which information is necessary
for your application will depend on the answer to the question whether you want to perform a full calibration,
whether you have a special setup or you have calibrated before and therefore want to take advantage of the user-
defined calibration. Furthermore, information about the calibration plate and the camera is required.
186 HDevelop Assistants
Choosing the Task for Your Application If you want to calibrate all parameters, for example, if you are cal-
ibrating for the first time with your setup, click the radio button Full Calibration: Pose and all
Camera Parameters.
If you are using a special setting or you have already calibrated your system before and want to re-use your
resulting parameters, choose User-defined: Select Individual Parameters for Calibration.
After having decided on your calibration task, proceed with the details about your Calibration Plate.
Calibration Plate Parameters First, choose the description file for your calibration plate and the calibration plate
Thickness (in mm). The name of the description file indicates the size of the plate and the file extension
indicates the type of the calibration plate. In particular, the file ending .cpd is used for calibration plates with
hexagonally arranged marks and the file extension .descr is used for calibration plates with rectangularly
arranged marks. Note that calibration plates with hexagonally arranged marks usually have light marks on a
dark background. Nevertheless, also calibration plates with dark marks on a light background are available.
Then, you have to choose a description file with the filename ending "_dark_on_light". Calibration plates
with rectangularly arranged marks always have dark marks on a light background. With the parameter
Thickness, you can modify the position of the world coordinate system and the measurement plane, which
is located beneath the calibration plate surface by the value of Thickness in the Reference Image (page
190).
Then proceed to set the Camera Parameters.
Set Camera Parameters For setting up the camera parameters
• first choose the Camera Model,
• then set the parameters for a full calibration or
• set the parameters for a user-defined calibration (page 188).
It is also possible to import parameters from a file. If you should decide to do this, just click the Import
Parameters button.
When you have finished this last part of the Setup tab, proceed to the Calibration (page 188) tab.
Choose your Camera Model First, choose the Camera Model you are using:
• either Area Scan (Division),
• Area Scan (Polynomial)
• or Line Scan
Even though the camera model Area Scan (Division) typically returns good results for your application,
you can improve the accuracy and lower the error rate by using the Area Scan (Polynomial) camera
model. We therefore recommend for you to use the Area Scan (Polynomial) model if the Mean Error
on the Results (page 195) tab under Calibration Status (page 195) is too high. If you decide for the
Area Scan (Polynomial) model, it is especially important that you thoroughly cover the field of view
with calibration plate images and do not leave out the edges.
Note, that a higher Mean Error might be caused by inappropriate calibration images.
For area scan cameras, the following additional options may be specified:
If your camera has a tilt lens, tick the checkbox Tilt.
Then, the Projection Model must be specified (either Projective or Telecentric).
Telecentric tilt lens cameras are available in different variations, which must be specified in Projection
Model:
• Bilateral Telecentric
• Object Side Telecentric
• Image Side Telecentric
Set Parameters for Full Calibration If you choose a full calibration for an area scan camera, you must specify
approximate values for the camera parameters. These parameters depend on the camera type and are listed
in the following tables. Please have a look at the data sheet of your camera for suitable values. Information
about the focal length can be found on the lens itself.
7.2.2 How to Calibrate With the Calibration Assistant 187
Assistants
Cell width X X
Cell height X X
Focal length X X
Image plane distance X
Tilt angle X
Rotation angle X
Parameters for telecentric area scan cameras:
Parameter no tilt bilateral object side image side
Cell width X X X X
Cell height X X X X
Magnification X X X
Focal length X
Image plane distance X
Tilt angle X X X
Rotation angle X X X
Parameters for line scan cameras:
Parameter
Cell width X
Cell height X
Focal length X
Motion X X
Motion Y X
Motion Z X
188 HDevelop Assistants
Set Parameters for User-Defined Calibration In the user-defined mode for the area scan camera
(division), you can also choose Center Column (Cx), Center Row (Cy) and Kappa. The
area scan camera (polynomial) model allows you to also choose the lens distortion parameters for
radial distortion Radial 2nd Order (K1), Radial 4th Order (K2), Radial 6th Order (K3) and
the two parameters for tangential distortion Tangential 2nd Order (P1) and Tangential 2nd Order
(P2).
If you want to change the parameters Cell Width and Cell Height independently from each other, click
the chain button.
The main part of the calibration process consists of acquiring images of the calibration plate in different positions
and orientations relative to the vision system. Please note that the more you vary the position and orientation, the
better the calibration results will be. Good calibration images will improve your calibration results significantly.
Detailed instructions on how to take calibration images can be found in the section “How to take a set of suitable
images?” in the reference manual chapter “Calibration”.
Obligatory steps for calibration are
ATTENTION: Keep your camera setup (aperture, focus, pose) fixed when you have chosen it! This applies
to the calibration process itself as well as to the subsequent application. Any changes will result in the failure
of the calibration or, even worse, in wrong output values.
When the calibration images are available, you can push the Calibrate button and move on to the Results (page
195) tab.
Assistants
a) b) c)
d) e) f)
Figure 7.1: Acquisition of suitable calibration images using a calibration plate with hexagonally arranged marks. The
plate is a) placed in the measurement plane, b) - e) tilted in different directions, and c) placed parallel
to the measurement plane (with a rotation around the z axis).
Figure 7.2: Example for suitable calibration images using a calibration plate with hexagonally arranged marks.
If you use the Image Acquisition Assistant and want to see a live image, you can also activate Live Image
on the Calibration tab and click the Snap button whenever you want to keep an image for calibration.
If you Load... images from a file into the Calibration Assistant and then decide to acquire new images
with the Image Acquisition Assistant, you will be warned that the images from the file will be removed from
the window.
With the Remove and Remove All buttons on the left, you can remove either one or all images of the list.
The Save and Save All buttons will save one or all images of the list.
Click Update to control the time when camera parameters, segmentation parameters or quality adjustments
have to be transferred for the existing images.
190 HDevelop Assistants
a) b)
Figure 7.3: Images of calibration plates with their coordinate systems: a) plate with hexagonally arranged marks
and b) plate with rectangularly arranged marks.
Activate Auto Update to automatically update to the latest adjustments. Quality Issues are updated with
a little delay after adapting Calibration Plate Extraction Parameters (page 194). Deactivating
Auto Update enables you to change several parameters at once and speeds up the processing of bigger data
sets.
Select a Reference Image With the pose of the calibration plate in the reference image, you specify the world
coordinate system and the measurement plane for subsequent 3D measurements (see figure 7.3). Thus, in
one calibration image (typically, the first one), you should place the calibration plate such that it lies on top
of the measurement plane.
If this is not possible, place the calibration plate in a position parallel to the measurement plane and “move”
the coordinate system by adapting the parameter Thickness.
The star on the left side of the Calibration window indicates the reference image. It is by default set on
the first image. You can, however, by clicking the Set Reference button, pick another image as reference.
Calibrating
Click the button Calibrate to finally perform the calibration task.
Check first whether you have enough images of sufficient quality. You can check the quality under Quality
Issues.
If necessary, you can also change Calibration Plate Extraction Parameters (page 194) before ac-
tually calibrating. In case your calibration fails and displays the error "Camera calibration did not
converge", check possible error sources in the Checklist for Calibration Errors or have a look at the rec-
ommendations listed in the reference manual chapter “Calibration”, section “How to take a set of suitable
images?”.
7.2.2 How to Calibrate With the Calibration Assistant 191
Assistants
pose of the camera) change during the calibration settings during calibration and later during the ap-
process? plication. If you decide to change anything you
have to start a new calibration.
Did you acquire the calibration images the way they Make sure you followed the recommendations in
are required? the section “How to take a set of suitable images?”
in the reference manual chapter “Calibration” re-
garding the placement of the calibration plate, the
setup of the cameras and the image properties.
Are you using an extreme wide angle lens? The distortions that appear close to the image bor-
ders cause a higher Mean Error or can even be
responsible for the failure of the calibration. You
must use another lens in this case.
Is the size of your camera chip compatible with the Using a lens that is not compatible with your cam-
lens? era chip size (this information should be included in
the instructions of the lens) will decrease the quality
of your image.
Contrast is low The difference between the gray Reasons can be either overexpo-
values of the calibration plate and sure or underexposure. To im-
the calibration marks is not big prove your results, change your
enough. aperture or the brightness of your
illumination.
7.2.2 How to Calibrate With the Calibration Assistant 193
Diameter of marks are too small The diameters of the found cali- To fix this issue, you should either
Assistants
bration plate marks are too small. change your setup or use a calibra-
tion plate with larger marks.
Marks on plate are out of focus The marks are not completely The depth of field has to include
focused, some of them appear the whole object. To fix this error,
blurry. This leads to a lower ro- change either your focal length or
bustness. the distance of the object to the
camera. Alternatively you can
also make the aperture smaller and
use brighter illumination.
Quality assessment failed The image test failed, even though For calibration plates with rect-
the plate could be found in the im- angularly arranged marks, check,
age. if any part of the image is oc-
cluded and if the occlusion inter-
rupts the black margin of the cali-
bration plate.
Mark extraction failed for some It was impossible to extract the Delete the images for which mark
images calibration plate marks in some extraction has failed and use new
images, which makes it also im- images instead or adapt the extrac-
possible to calibrate in this state. tion parameters. Make sure that
you follow the recommendations
in the section “How to take a set of
suitable images?” in the reference
manual chapter “Calibration”.
194 HDevelop Assistants
Quality issues detected for some The quality of some images is be- Check the quality issues of the sin-
images low the warn level. gle images by clicking their names
in the list. Handle quality issues as
described in the table above.
Number of images is too low The number of images is lower Check if you have taken enough
than recommended. images, depending on the type of
calibration plate you use.
Field of view is not covered by Some part of the field of view is Press the Show button, which
plate images not covered by any image of the appears in a column named
calibration plate, that is there are Details, to see all areas in red
areas with no marks. that are not covered by calibration
plate images (compare figure 7.4).
Before continuing, add the miss-
ing image(s) to your sequence.
You can avoid this problem by
following the recommendations in
the section “How to take a set of
suitable images?” in the reference
manual chapter “Calibration”.
Tilt angles are not covered by se- The calibration plate has not been Add more images of your calibra-
quence tilted enough. tion plate tilted in different direc-
tions. For calibration plates with
rectangularly arranged marks we
recommend tilting the plate in ev-
ery quadrant of the image twice
and vary the tilting direction.
Not all image sizes are identical The image list contains images of You have changed your setup
different sizes. while taking calibration images.
Therefore, you should delete those
images taken before the change to
get useful results back.
In the drop-down menus under Display Parameters you can choose your own values for Plate Region, Mark
Centers or the Coordinate System. The Draw option lets you choose whether you want to see margins or filled
regions.
Some of the parameters under Calibration Plate Extraction Parameters may be changed to improve the
calibration results.
For calibration plates with hexagonally arranged marks:
• Smoothing (Sigma) should be set to a higher value if the image is blurry or contains strong noise.
• Gap Tolerance should be set to a higher value if the plate is strongly tilted and
• Smoothing (Alpha) should be set to a smaller value if the image is blurry.
• Furthermore the Maximum Mark Diameter may be changed if the checkbox is activated.
7.2.3 Results of the Calibration 195
Assistants
a) b)
c)
not covered
covered
d)
Figure 7.4: Output when not enough calibration images have been taken. .
For more information about these parameters, please refer to the reference manual entry of the HALCON operator
find_calib_object.
Two types of parameters of your vision system are calculated as results: internal parameters, for example, the
precise focal length, the size of the camera chip, or the distortion caused by an imperfect lens, and external
parameters, for example, the position and orientation of the vision system.
• Calibration Status,
• Camera Parameters, and
• the Camera Pose.
The results displayed in Camera Parameters and Camera Pose can also be saved to a file by clicking the Save
buttons on the right.
Display Results enables you to choose which results should be displayed.
When you are finished with the results, go on to the Code Generation (page 199) tab.
This box displays whether the calibration was successful, and the Mean Error in pixels.
If you either delete calibration images (page 188), change Calibration Plate Extraction Parameters
(page 194) or Camera Parameters (page 186) after having calibrated, the former calibration data is not valid
any more. Therefore, the Status will display that no calibration data is available. To continue working with
your changed camera parameters, calibration parameters or images, just click Calibrate (page 190) again on the
Calibration tab.
196 HDevelop Assistants
Mean Error Designates the average error in pixels during the calibration process. When the system has been
calibrated, the ideal centers of the calibration marks are calculated and compared to the real mark centers.
Mean Error is the deviation value between the ideal and the real mark centers. A value of 0.1 and lower
can be regarded as a good result. Possible calibration errors are described in the tables about quality issues
under Quality Issues (page 191).
The internal camera parameters include Cell Width (Sx) and Cell Height (Sy) in µm, Focal Length in
mm, Center Column (Cx) and Center Row (Cy), Image Width and Image Height in pixels. They also
include Kappa in m−2 or instead of Kappa, the distortion parameters Radial 2nd Order (K1) in m−2 , Radial
4th Order (K2) in m−4 , Radial 6th Order in m−6 , Tangential 2nd Order (P1) and Tangential 2nd
Order (P2) in m−2 for the polynomial area scan camera model.
If you have a line scan camera, additionally to the values of the area scan camera (division) model, values
for the motion parameters Motion x (Vx), Motion y (Vy) and Motion z (Vz) in µ/pixel will be returned.
The 3D pose of the world coordinate system relative to the camera is described by the external camera parameters
X, Y and Z in mm and Rotation X, Rotation Y and Rotation Z in degrees.
You can choose Original Reference Image to see the previously chosen reference image and Simulated
Reference Image to display a synthetic reference image, which has been calculated using the internally known
measures of the calibration plate and the pose of the plate in the reference image, using the radio buttons. You can
also decide whether you want to Display Coordinate Axes of the coordinate system of the calibration plate.
This tab helps you to generate and insert code for calibrating and for using the calibration results in your HDevelop
program. The tab is subdivided into four parts:
• Calibration
• Sample Usage
• Variable Names (page 198)
• Code Preview (page 198)
When you are finished with configuring the options, check the position of the insert cursor and click Insert Code
(page 199) under Calibration or Sample Usage to insert the code.
! If you have already inserted code into your program and you click insert code again, the previous code will be
replaced regardless of the cursor position.
7.2.4.1 Calibration
For the last one you can click the directory icons to browse for a stored file. Subsequently select Parameter File
and Pose File.
Assistants
In order to save the calibration results to files it is necessary that
• a successful calibration took place before and
• a file name exists for both files.
To generate code for initializing the image acquisition when using the Image Acquisition Assistant (page
180), enable Initialize Acquisition.
When you are finished, check the position of the insert cursor and click Insert Code (page 199) to insert the
code into your HDevelop program.
The Browse button on the Code Generation tab is similar to the Save button on the Results tab. It can be used
to create file names into which the calibration results can then be written, when choosing the option Calibration
Results (File).
Sample Usage shows you what is possible with your calibration data and provides code, which you can adapt
to your own purposes. Choose the action you are interested in and the example code will be inserted into your
program.
You have the choice between:
When you are finished, check the position of the insert cursor and click Insert Code (page 199) to insert the
code into your HDevelop program.
For each calibration, default variable names are chosen. You can, however, use your own variable names and
change variable names for:
• Connection Handle
• Image Object
• Camera Parameters
• Start Parameters
• Loop Counter
• Image Files
• Camera Pose
• Window
Note: These are variables which you might set before the generated code or use after the generated code. Interme-
diate variables have fixed names starting with TmpCtrl or TmpObj.
When you are finished, check the position of the insert cursor and click Insert Code to insert the code into your
HDevelop program.
Here, you can, for example, edit or replace individual operators of the code lines proposed by the Calibration
Assistant.
For details, see also Code Generation (page 200) in the menu.
Assistants
• Calibrate (page 188)
• Help
Tabs with the dialogs for most of the tasks that can be done with the Calibration Assistant:
• Setup
• Calibration
• Results
• Code Generation (page 201)
Furthermore, it provides a status bar at the bottom in which messages are displayed. The status bar also displays
the calibration results (page 195), that is, if the calibration was successful. Please note that the status bar does
not provide a scrolling mechanism; if the displayed message is too long, move the mouse over it, so that a tool tip
displaying the full message pops up. Alternatively, if the message is only slightly larger than the status bar, you
can also drag the left or right border of the Calibration Assistant window to enlarge it.
Images are displayed in the graphics window of HDevelop.
Via the menu Calibrate you can run a calibration as described in the section Calibrating (page 190).
Via the menu Help you can access the online documentation.
Assistants
The Code Generation tab includes the following subdivisions:
Matching Matching is the process of locating an object described by a model in an image. The main results of
the matching process are the position and orientation of the found object instance and its matching score.
Alignment This method can be applied to transform the position of the matched object corresponding to the
reference image. Alignment is useful if the following image processing step is not invariant against rotation
or translation, like OCR or the variation model. Note that by alignment the matched object is only rotated
and translated. More information about alignment using shape-based matching can be found in the Solution
Guide II-B, section 2.5.3 on page 33. To remove perspective or lens distortions, for example, if the camera
observes the scene under an inclined angle, you must rectify the image first.
Model The model is an internal representation of the object containing only the information characterizing the
object. This model is created by the assistant from an example image of the object, the model image, which
is provided by you. The model is used when searching for the object in the test images. You can also provide
the Matching Assistant with a model, see section 7.3.3.1 on page 204.
Model Image This is the image containing your example of the object to be searched for. This image should be
a characteristic image of the object. The object should appear in its default position and orientation and not
be occluded.
Reference Image If a reference image is selected, the position of the match in this image is used as reference.
This is necessary to perform alignment. If no reference image is chosen, the model image will be used as
basis for alignment.
Model Region of Interest (ROI) This is the region in the model image which contains the object to be trained.
You can select this region via the menu item ROI (page 205).
Rectification This method can be applied to the search image to transform it such that the found model and the
model in the reference image appear as similar as possible. Rectfying an image is useful for all further
processes that rely on fixed ROIs, like measuring and OCR.
Test Image You can test the performance of the matching process by providing test images. These images should
be representative images from your matching application. The object should appear in all allowed variations
of its position, orientation, occlusion, and illumination.
7.3.2 How to Use the Matching Assistant of HDevelop 203
Assistants
By using the Matching Assistant, you can set up and optimize your matching application in four steps.
We recommend resetting all model and search parameters via the toolbar button Reset Model (page 207) before
starting with a new matching application. This way, the parameters are reset to their default settings and the model
image, the model ROI, and the test images are deleted.
Select the matching method from the pull-down menu in the toolbar (page 207) of the Matching Assistant.
(1) Load the model image. This can be done via the menu item File . Load Model Image or the corresponding
button and text field under Model and Model Source inside the tab Creation (page 208).
Alternatively choose the Image Acquisition Assistant (page 180) as image source.
(2) Specify an ROI around the object. This can be done via the menu item ROI (page 205) or the corresponding
buttons inside the tab Creation (page 208).
(3) Specify standard and advanced model parameter values that are available for your matching method. This can
be done via the menu Parameters (page 206) or the tab Parameters (page 210).
• Load one or more test images. This can be done via the menu item Usage . Test Images . Load Test
Images (page 206) or the button Load in the tab Usage (page 214).
Alternatively chose the Image Acquisition Assistant (page 180) to acquire images by activating the corre-
sponding checkbox under Test Image Source.
• Specify standard search parameter values. This can be done via the menu item Usage . Standard Model
Use Parameters (page 206), or the menu item Standard Use Parameters (page 215) in the tab Usage
(page 214).
• Assure that all objects are found in all test images by comparing the number of existing model instances
with the number of found instances (applying for example, Find Model in Usage (page 214)) or simply
determine the recognition rate (page 218).
Configure your matching process such that the search is successful in all test images. To do so, you can start to
optimize the parameters on the tabs Parameters and Usage.
You can also optimize your application regarding the speed. How to do so depends on the matching method and
the specific application. We recommend to check following parameters which we see as especially promising:
• shape-based matching: The Matching Assistant allows you to optimize the search parameters Minimum
Score (page 215) and Greediness (page 215). See Solution Guide II-B, section 3.2.9 on page 73.
• deformable matching: The Matching Assistant allows you to optimize the search parameters Minimum
Score (page 215) and Greediness (page 215). See Solution Guide II-B, section 3.3.4 on page 82.
204 HDevelop Assistants
• correlation-based matching: The search parameter Minimum Score (page 215) can be tuned. See Solution
Guide II-B, section 3.1.4 on page 50.
• descriptor-based matching: The model parameters Fern Number (page 210) and Fern Depth (page 210)
can be tuned. Few ferns with a large depth enable a fast online matching which, however, requires more
memory. See Solution Guide II-B, section 3.5.4 on page 99 for further optimization possibilities..
To find out if a speed up is necessary or successful, it is useful to check the Statistics on the tab Inspect (page
218).
Search parameters can also be automatically improved via the menu item Usage . Go To Optimize
Recognition Speed (page 206), or the menu item Optimize Recognition Speed (page 217) in the tab Usage
(page 214). But we recommend checking whether the matching still succeeds in all test images after each modifi-
cation.
Insert the code performing the chosen matching tasks within an HDevelop program. This code includes all of the
previous adaptions and can be used in the final application.
The menu Code Generation (page 207) and the tab Code Generation (page 218) allow you to choose between
several options and change parameter names, which has a direct effect on the generated code.
The Matching Assistant shown in figure 7.5 on page 209 consists of the following elements:
• The following pull-down menus: File, ROI, Parameters (page 206), Usage (page 206), Inspect (page
207), Code Generation (page 207), and Help (page 207)
• A tool bar (page 207) for direct functionality access.
• The following tabs: Creation (page 208), Parameters (page 210), Usage (page 214), Inspect (page
218), and Code Generation (page 218)
• A status bar at the bottom in which messages are displayed. The status bar also displays the matching results.
The number of found instances, the needed time, and for each found instance the position, orientation, scale,
and score. Please note that the status bar does not provide a scrolling mechanism; if the displayed message
is too long, move the mouse over it, so that a tool tip displaying the full message pops up.
Note, not all options and parameters are supported for every matching method.
Assistants
Parameters
Restriction: Only supported for descriptor-based matching and deformable matching.
Load a camera pose from file.
Load Camera For more information on poses, see for example, “Transformation . Poses” and Solu-
Pose tion Guide III-C, section 2.1.4 on page 20.
Restriction: Only supported for descriptor-based matching and deformable matching.
Display Image Inspect the model image pyramid and open the corresponding dialog in the tab
Pyramid Level Creation (page 209).
Load the settings of a previously saved Matching Assistant session.
Note: If the settings file refers to a model image file that is no longer available because
Load Assistant
it has been moved or deleted since, you can choose to select an alternate model image.
Settings
If (some of) the test images cannot be loaded, a message box with the missing image
file names is displayed.
Save Current
Assistant Save the current settings of a Matching Assistant session.
Settings
Closes the Matching Assistant dialog.
In doing so, the current settings are stored for the duration of the current HDevelop
Close Dialog
session. That is, as long as you do not exit HDevelop, you can reopen the Matching
Assistant with the same settings.
Close the assistant’s dialog and exit the Matching Assistant without storing the current
Exit Assistant
settings.
The menu ROI allows you to generate the model region of interest by drawing and modifying it on the displayed
model image.
To draw objects, click into the image and move the mouse over the object while keeping the left mouse button
pressed; the selected shape appears. After releasing the mouse button you can move the ROI by dragging its center
(marked with a cross) with the left mouse button. Furthermore, you can edit the shape by dragging its boundaries.
Finish the creation by clicking once with the right mouse button.
The menu Parameters allows you to adapt model parameters suitable to your matching application.
Note: Depending on the matching method, different parameters are available for adaption.
The menu Usage allows you to apply your model on test images and adapt suitable search parameters.
Assistants
Recognition
214).
Speed
The menu Inspect allows you to determine the recognition rate and the pose bounds of the found object instances
for the used set of test images.
The menu Code Generation allows you to generate and modify the HDevelop code for your specified matching
application.
The menu Help allows you to open the HALCON help window.
The tab Creation allows you to create or load a matching model. It consists of different sections.
Model
Determine the source type for your model.
Model Source
Determine the source for your model.
Model ROI
7.3.3 Matching Assistant Reference 209
Assistants
Figure 7.5: The matching assistant and its different elements.
Pull-down menus
Tool bar
Tabs
Status bar
Depending on the source type, you can draw an ROI. Next to the options explained in the The Menu ROI
(page 205), you have the following option.
on the size of the model ROI; depending on the selected Contrast and Min. Component Size, higher
pyramid levels may not contain any model points.
Calibration Source
Select a calibration source.
Assistants
be included in the model.
Min. Component
For information on the functionality of the parameter see its documentation in
Size
set_generic_shape_model_param.
Auto Select: The Matching Assistant selects a suitable value automatically
based on the model image.
SBM: 3, CBM: 3, DBM: 5, DM: 3
Select the number of pyramid levels. You can enter the value directly in the text field
or by using the slider next to it.
See the Solution Guide II-B, section 2.3 on page 25 for more information about the
image pyramid and its levels.
See also Display Image Pyramid (page 209) for information how to inspect the
Pyramid Levels model image pyramid.
Auto Select: The Matching Assistant selects a suitable value automatically
based on the model image.
Please note that in rare cases the automatic selection will yield a too low value and
thereby slow down the search process, or a too high value, resulting in failures to rec-
ognize the object. In such a case we recommend that you inspect the model image
pyramid (page 209) and select a suitable value manually.
SBM: 5, CBM: 3, DBM: 5, DM: 3
Specify the starting angle of the allowed range of rotation (unit: °).
Example: To allow model rotations up to +/-5 °, for example, you should set the start-
Starting Angle ing angle to -5 ° and the angle extent to 10 ° or angle end to 5 °.
Note: The range of rotation is defined relative to the created model, that is, for a model
created from an image a starting angle of 0 ° corresponds to the orientation the object
has in the model image.
SBM: 5, CBM: 3, DBM: 5, DM: 3
Specify how much the object is allowed to rotate (unit: °).
Angle Extent
Example: To allow model rotations of up to 10 °, for example, you should set the angle
extent to 10 °.
SBM: 5, CBM: 5, DBM: 3, DM: 5
Define the range for the angle of rotation around the normal vector of the model.
Min./Max. Angle
For further information see create_uncalib_descriptor_model, part “Simulation
parameters” ’min_rot’ and ’max_rot’.
SBM: 3, CBM: 5, DBM: 5, DM: 3
Define the allowed scaling range.
The allowed range of scale is defined separately in row and column direction. Thus, it
is described by the parameters:
• Min. Column Scale
• Max. Column Scale
• Min. Row Scale
• Max. Row Scale
Min./Max. The model template has scaling values of 1.0. This means, if the model was generated
Column/Row Scale from a model image, the scaling values of the model instance in this image are 1.0.
Depending on the specified parameters, the most efficient scaling method is used. This
means:
• Unscaled matching: This method is used if all four scale factors are equal to
1.0.
• Isotropically scaled matching: This method is used if rows and columns have
the same scaling factors (but not all are equal to 1.0).
• Anisotropically scaled matching: This method is used if rows and columns
have different scaling factors.
212 HDevelop Assistants
Assistants
For further information see Solution Guide II-B, section 3.2.4.4 on page 61
(shape-based matching), create_ncc_model (correlation-based matching), and
Angle Step find_planar_uncalib_deformable_model (deformable matching).
Note: Each time you create a model ROI or change the parameter Contrast (page
210), the Matching Assistant automatically selects a suitable value for Angle Step
to obtain the highest possible accuracy.
Auto Select: The Matching Assistant selects a suitable value automatically.
SBM: 3, CBM: 5, DBM: 5, DM: 3
Determine the step length within the selected scaling in row and column direction.
For further information see Solution Guide II-B, section 3.2.4.3 on page 60 (shape-
based matching) and find_planar_uncalib_deformable_model (deformable
Row/Column Scale
matching).
Step
Note: Each time you create a model ROI or change the parameter Contrast (page
210), the Matching Assistant automatically selects a suitable value for Row Scale
Step and Column Scale Step to obtain the highest possible accuracy.
Auto Select: The Matching Assistant selects a suitable value automatically.
SBM: 3, CBM: 3, DBM: 5, DM: 3
Specify how pixels are compared with the model, more precisely whether and how the
polarity must be observed.
Metric For information on the functionality of the parameter and its possible values see So-
lution Guide II-B, section 3.2.4.5 on page 62 (shape-based matching) Solution Guide
II-B, section 3.1.3.3 on page 49 (correlation-based matching), and Solution Guide
II-B, section 3.3.3.4 on page 81 (deformable matching).
SBM: 3, CBM: 5, DBM: 5, DM: 3
Determine the number of points by which the model is reduced.
For further information see Solution Guide II-B, section 3.2.4.2 on page 59 (shape-
Optimization based matching) and Solution Guide II-B, section 3.3.3.2 on page 81 (deformable
matching).
Auto Select: The Matching Assistant selects a suitable value automatically
based on the model image.
SBM: 3, CBM: 5, DBM: 5, DM: 3
Determine the minimal contrast (gray value difference to neighboring pixels) a point
in a search image must at least have in order to be compared with the model during
the matching process.
Min. Contrast For information on the functionality of the parameter see
set_generic_shape_model_param (shape-based matching) and Solution Guide
II-B, section 3.3.3.4 on page 81 (deformable matching).
Auto Select: The Matching Assistant selects a suitable value automatically
based on the model image.
SBM: 5, CBM: 5, DBM: 3, DM: 5
Determine the number of checked neighbors in the circle when using the Lepetit
method.
Check Neighbor
For further information see create_uncalib_descriptor_model and
CheckNeighbor in points_lepetit.
Restriction: Only for detectors of type lepetit.
SBM: 5, CBM: 5, DBM: 3, DM: 5
Determine the threshold of gray value difference to each circle point when using the
Neighbor Diff. Lepetit method.
Threshold For further information see create_uncalib_descriptor_model and
MinCheckNeighborDiff in points_lepetit.
Restriction: Only for detectors of type lepetit.
SBM: 5, CBM: 5, DBM: 3, DM: 5
Turn the refinement for subpixel accuracy on or off.
Subpixel For further information see create_uncalib_descriptor_model and SubPixel
in points_lepetit and points_harris_binomial.
Restriction: Only for detectors of type lepetit and harris_binomial.
214 HDevelop Assistants
The tab Usage allows you to apply your model on test images and adapt suitable search parameters.
Test Images
You can select a test image from the text box of the dialog Test Images. The selected image is automati-
cally displayed in the graphics window of HDevelop.
Assistants
Save All Save all images added by the Image Acquisition Assistant (page 180).
Set the match in the chosen image as reference position for alignment.
Set Reference Note: This is only possible if a match is detected in the image. If no reference image
is chosen, the model image will be used as basis for the alignment or rectfication.
Search all objects in the complete sequence of test images that were loaded before
(independent of the set Maximum Number of Matches). The results are displayed
successively in the graphics window.
Note: For descriptor-based matching this also sets the Maximum Number of
Detect All Matches if it has not been set previously (it is set to 0).
Note: This also sets the number of visible objects in the test image if it has not been
set previously.
Note: If the button is clicked for the first time or after you changed a model parameter,
the internally stored model is actually created, which takes some time.
Search the object in the currently selected test image. The result is displayed in the
assistant window.
Find Model
Note: If the button is clicked for the first time or after you changed a model parameter,
the internally stored model is actually created, which takes some time.
Start the matching process automatically on the selected test image. The result is
displayed in the graphics window.
Always Find
Note: If the matching process is started for the first time of after you changed a model
parameter, the internally stored model is actually created, which takes some time.
Specify how many objects are really visible in the current test image. To do so, click
Number of Visible onto the currently displayed number of detected objects (Visible) in the text field of
Objects in Test Im- the currently selected test image. The specified numbers of visible objects are used
age when determining the recognition rate, that is, the recognition rate is 100% when the
sum of all objects found in the test images is equal to the sum of the specified numbers.
Assistants
Descriptor Min.
of the model in the image.
Score
For more information on the parameter see Solution Guide II-B, section 3.5.4.4 on
page 101.
SBM: 5, CBM: 5, DBM: 3, DM: 5
Enable the usage of an already approximately known homography to ’guide’ the
Guided Matching matching process, which enhances the accuracy of the object recognition.
For more information on the parameter see Solution Guide II-B, section 3.5.4.3 on
page 100.
The tab Inspect allows you to obtain the Recognition Rate and Statistics.
Recognition Rate The recognition rate calculated for different criteria. Note that it depends on the set Maximum
Number of Matches (page 215) and specified number of visible objects (page 214).
Statistics The minimum, maximum and extent for different pose parameters of the found instances as well as
for the score and execution time are shown. The range of positions, orientations, and scales in which the
object appears in the test images are shown.
The tab Code Generation allows you to generate and modify the HDevelop code for your specified matching
application.
Options
This section allows you to determine, which code parts are to be inserted.
Load model file: In the application code the matching model will be loaded from file.
Load recently
In the application code the recently saved model will be loaded from file.
saved model file
Initialize
Insert code to initializing an image acquisition device.
Acquisition
Generate display
Display the detected model instances in a loop.
code
Generate value
Generate code to retrieve the parameter values of the found matches in a loop.
retrieving code
Note: For shape-based matching only.
Insert code for an affine transformation that can be used for alignment.
Generate affine
Note: A reference image (page 214) has to be set on the Usage tab, otherwise the
transformation
model image is used.
code
Restriction: Only for shape-based matching.
Generate Insert code for rectifying the image of detected model instances.
rectification Note: A reference image (page 214) has to be set on the Usage tab, otherwise the
code model image is used.
This section allows you to modify the names of individual variables needed for the code lines.
The dialog consists of a text field for every variable. The Matching Assistant automatically generates rea-
Assistants
sonable variable names, but you can change the individual names via the text fields.
Code Preview
This section allows you to preview the generated code and gives you the possibility to, for example, edit or
replace individual operators of the code lines proposed by the Matching Assistant.
The Measure Assistant of HDevelop is a front-end to HALCON’s 1D measuring. Using the Measure Assistant you
can, for example,
• easily set parameters and perform a visual inspection,
• quickly measure distances between edges along a straight line or circular arc,
• choose the results you need (page 222),
• perform a calibration to transfer your results to world coordinates (page 223), and
• if necessary, use advanced features to, for example, optimize your measuring under difficult conditions with
fuzzy measuring (page 225).
Using the Measure Assistant is simple: all you need is to setup the measure task to detect all relevant edges and
add the resulting code to your application.
When you start a project for the first time, read “How to Use the Measure Assistant of HDevelop”.
When looking for an overview over all Measure Assistant elements, please refer to the reference (page 228).
In this online help, the following special terms are used:
The Measure Assistant allows you to easily choose the dimensions of an object you need in just a few steps.
It is very easy to set up the measure task with the Measure Assistant:
• Choose your Image Source on the Input tab by either using the content of the Graphics Window, loading
an Image File, or acquiring the image of an object using the Image Acquisition Assistant.
• Create an ROI using the buttons Draw Line or Draw Circular Arc, which you find in the toolbar of the
Measure Assistant. Place the ROI perpendicular to the edges you want to measure.
• Now, that you have located the ROIs, you can extract edges within the chosen region(s) and therefore use the
functionality of the Edges tab.
The Input tab offers two more options that you can use if necessary. If you already have calibration data available,
you can load the data as described in the paragraph Calibration Source (page 223), otherwise just return to the
Input tab later. A more advanced feature is changing the mode of the expected gray-value range which is only
relevant if your image data uses more than 8 bits. Read more about this topic under Expected Gray Value
Range (page 224).
In case your edge extraction is not as successful as expected, further improvements can be made using the fuzzy
measuring function of the assistant (page 225).
Image Source
The Input tab lets you choose images from the following sources:
• If your image has already been opened in the Graphics Window, you can activate Graphics Window
to continue working with the currently displayed image.
• You can load an image file by activating Image File or choosing Load Image from the menu File
and the tool bar, respectively, and either typing in the image path or using the Browse button on the
right to choose an image from a file.
• Another option is to activate the radio button for the Image Acquisition Assistant. Being connected
to this assistant, you can acquire the image you want to measure in and you can even choose to work
with a Live image. The same operations can be performed by clicking the Snap or Live symbols in
the tool bar.
Note that the measuring works on a single channel. For color RGB images, the red channel will be used. A
color transformation can be performed with the operator trans_from_rgb.
When you have loaded your image, you can continue to create an ROI. If your image data exceeds 8 bit, you
can change the expected gray-value range (page 224). If you want to load calibration data or calibrate now,
proceed to the section Calibration Source (page 223).
Create an ROI
Create an ROI by using the buttons Draw Line or Draw Circular Arc in the tool bar. The shape of the
ROI should be chosen according to the shape of the object to be inspected.
After having chosen the ROI shape, you “draw” the ROI in the Graphics Window by keeping the left mouse
button pressed. Then, modify the ROI until it has the correct shape. For linear ROIs you can modify the
length by “dragging” the line’s end points or move it by dragging its center. For circular ROIs you can
additionally modify the radius by dragging one of the four circle points that are located at 0, 90, 180, and
270 degrees. When you are finished, click the right mouse button, to confirm your choice. The ROI will
already display edges if those can be detected with default parameters.
Remember that the ROI should run perpendicular to the edges you want to measure.
If you have previously prepared and saved an ROI, you can reuse it by choosing Load ROI.
In order to delete one or more ROIs you can either mark them and just press the delete button on your
keybord or click the tool bar buttons for Delete Selected ROI Item or Delete All ROIs.
7.4.2 How to Use the Measure Assistant of HDevelop 221
You can view and edit the ROI data, which includes all the data about the exact position of the ROI(s), using
the tool bar button View ROI Data. Editing an ROI by changing the ROI data is useful if an ROI should be
Assistants
modified more precisely than is possible by drawing in the graphics window.
When you are satisfied with the shape of your ROI, proceed to the edge extraction on the Edges tab.
After you have prepared the measure task (page 220), you continue to choose parameters on the Edges tab such
that you can detect the edges between which you want to measure:
• On the tab Edge under Edge Extraction you can specify the parameters that are used to extract edges.
• Edge Selection allows you to group edges to pairs or choose edges with certain features.
• In order to improve the visibility of your ROI and edges, you can change the display parameters.
Now, you can select edges you want to measure and afterwards proceed to the Results tab to view your measuring
results.
Open the Line Profile window (page 107) to inspect edges along an ROI. If necessary, that is, if your edge
extraction is not as successful as expected, you can refine the determination of the edges you need for measuring
by using the Fuzzy Measure (page 225) option you find on the Fuzzy tab. Fuzzy measuring allows you to specify
certain ranges within which edges are labeled “good”.
Edge Extraction
You can optimize edge extraction by adapting the following parameters on the Edges tab:
• With Min. Edge Amplitude you can specify how strong an edge must be to be selected. Adapt this
parameter such that only your desired edges are selected. Note that very small values should only be
used with high quality images, otherwise false edges might be caused by noise.
• Smoothing is helpful to reduce noise and therefore the detection of false edges. Please note, however,
that smoothing distorts the edge profiles, i.e, edges are detected at a slightly wrong positions and the
accuracy decreases. When using the line profile tool (page 107), smoothing can either be applied with
the Measure Assistant or it can be performed with the smoothing slider within the line profile window.
The smoothing values are immediately transfered to the Measure Assistant and the line profile window,
respectively.
• With ROI Width you can specify how many pixels perpendicular to the line or arc of measuring are
used to detect edges. A larger value helps to reduce noise; however, if the edges themselves are not
perpendicular to the line or arc of measure you must choose a smaller value, otherwise the edge will
“loose its strength”. The general rule here is to choose the ROI always perpendicular and as wide as
possible. This will lead to both a high precision and accuracy. If changing the ROI width leads to an
edge length that is inconvenient for the display, for example, because it is very short or very long, you
can choose another length for the display instead of the ROI width under Display Parameters.
• With the Interpolation Method you influence the accuracy of the edge extraction, but also the pro-
cessing time. Processing time and accuracy both increase from nearest_neighbor over bilinear
up to bicubic. Activating Highest Accuracy influences the way the interpolation is calculated, it
takes a bit more processing time and also slightly improves accuracy.
If one or more edges cannot be found without further processing, or if false edges are detected, the Line
Profile (page 107) which can be opened with the View Line Profile button on the upper right edge of
the Edges tab can provide the missing details to solve the problem. In its display, the line profile shows a
green line with a gray top for each successfully detected edge that is above a certain level to qualify as an
edge. Potential edges that do not qualify as edges are displayed as vertical red lines. It may also be helpful
to view the line profile along an ROI while changing parameters for edge detection, to see exactly how your
actions influence the line profile and therefore possibly change which edges are detected. This profile can
consequently be a basis for the decision which steps should be taken next, that is, what would be necessary
to enhance weak edges or suppress false edges.
In order to select which edges or edge pairs from your current results are relevant, proceed to Edge
Selection. If you are already satisfied with the detected edges, continue to the Results tab.
222 HDevelop Assistants
Edge Selection
Edge Selection allows you to specify what kind of edges you are looking for. First you decide whether
you want to find single edges or edge pairs. Depending on your choice, you have different options for
the drop-down menus Transition and Position. These options will be explained for edge pairs in the
following paragraph and subsequently for single edges.
You can activate Group Edges to Pairs to find an even number of edges with alternating transitions. The
parameter Transition lets you select which edges are chosen to resolve ambiguity: The default setting of
all will always accept the first matching edge. The values positive and negative restrict the transition
of the very first edge. The *_strongest variants will pick the strongest edge instead of the first edge in a
sequence of edges with the same transition. Position lets you decide whether you are looking for all edge
pairs, or just want to detect the first or last edge pair.
If you are looking for single edges (instead of edge pairs) you can choose with Transition whether you
want edges with all transitions (positive and negative) or just positive or just negative. Using
Position, you can then determine whether you are looking for all edges or just want to detect the first
or the last edge.
When you are satisfied with your selection, you can view your measuring results on the Results tab.
Changing display parameters on the Edges tab may help you to make the ROI in your image better visible and
therefore easier to work with.
You can change
• the Region Color and also activate Show Region to see the actual size of your region which may be
convenient if you do not use the ROI width as edge length.
• the Edge Color and also activate Use Shadows if this is beneficial for the visibility.
• the Edge Length, which is the ROI width in default mode, to improve the visibility of the edges, and
• the Line Width.
When you are satisfied with your detected edges as well as with the display parameters, continue to the Results
tab.
7.4.3 Results
The Results tab lets you choose which results are relevant for your application and displays these results in a
spreadsheet.
When you are satisfied with the results, continue to the Code Generation tab to receive the code for your appli-
cation. Otherwise, go back to adapt your edge extraction (page 221) or calibrate your system.
Feature selection lets you choose the features that are relevant for your application. The results of those features
are measured and displayed below. All features are activated by default. It is, however, useful to deactivate those
which are not needed to improve performance and readability.
You can choose between:
• Position,
• Amplitude,
• Distance,
• Pair Width, and
• Fuzzy Score.
When you have chosen your features, you can either continue to Feature Processing if you want your results
transformed into world coordinates, or, proceed to Edge Data to view the results for each ROI.
7.4.4 Code Generation 223
Assistants
Here, you can decide whether you want to receive your results in world coordinates and if so, you can choose a
suitable Unit. This step will be unavailable if you have not calibrated your setup. If you want to transform your
measurement results into world coordinates, you should now choose a calibration source or calibrate your camera
system and load the data into the Measure Assistant (page 184) and then go back to Transform Image into
World Coordinates under Feature Processing.
Calibration Source
If you need results in world coordinates, a calibration is necessary. The Input tab lets you load calibration
data if available or opens the Calibration Assistant to calibrate live.
Choosing Calibration Files allows you to load calibration data from file which is useful if you have
already performed a calibration for your application. You can either type in the path to the parameter names
(*.cal) and the camera pose (*.dat) or use the Browse buttons next to the input boxes to load them from file.
Choosing Calibration Assistant (page 184) allows you to use calibration data from a calibration assistant
that quickly guides you through a calibration. A new assistant will be opened unless there is already one
available.
If you are just finished with your Input (page 220), continue by creating an ROI (page 220). If you have
already extracted your edges, proceed to examine your results under Edge Data.
Edge Data displays the measurement results. In order to see the different results for multiple ROIs, choose the
ROI you want to view the results of by clicking the name of that ROI in the Active ROI field. The results can
then be examined below.
If you want to use measurement results in other documents, you can simply mark entries, use Ctrl+C to copy those
entries to the global clipboard and subsequently paste them into any other document.
When you are satisfied with the measuring results, proceed to the Code Generation tab.
Code Generation produces the code that is necessary to perform the chosen measurement tasks within an HDe-
velop program. On the Code Generation tab you can choose between several options and change parameter
names, which has a direct effect on the code that is generated.
First, you can choose to Initialize Acquisition or Initialize Calibration automatically in your code.
Depending on whether you have previously used the Image Acquisition Assistant or the Calibration Assistant to
acquire images or get calibration data, respectively, you can now decide whether you want the generated code
from these assistants integrated in your code. Such an automatic integration of code results, for example, in the
automatic opening of the framegrabber if Image Acquisition Assistant is activated. If code for image acquisition
or calibration already exist, the checkboxes should be deactivated. If Image Acquisition Assistant or Calibration
Assistant are not in use, Initialize Acquisition and Initialize Calibration are grayed out.
Under General Options, you can decide on the Alignment Method, that is, whether an alignment is needed for
your application. To learn more about this advanced setting, please refer to the description in the section Alignment
(page 228).
You can furthermore change variable names for
• General parameters,
• ROI coordinates, and
• Measurement results.
When you are finished changing parameters, click the Insert Code button under Measuring to generate the
code.
Finally, integrate the code into your HDevelop program.
224 HDevelop Assistants
If desired, you can change the default variable names for general parameters, ROI coordinates and measurement
results or replace them with your own variable names.
When you are finished changing your variable names, proceed to the Code Preview.
Before clicking the Insert button to include your code into the program window, you can preview the code in the
Code Preview table. You can now step through the table which consists of the columns
• Insert Operator, which shows the operator that will be inserted when you press the Insert button,
• Procedure, which shows the corresponding procedure,
• Line, referring to the line number within the code, and
• Replace Operator which shows previously generated code that will be replaced.
This section deals with more complicated measuring tasks. Therefore, you might find this section useful
If your image data exceeds 8 bits, it might be useful to choose a minimum and maximum gray value and change
the mode that handles the expected gray-value range. On the Input tab, there are three modes available that allow
you to control how much adaption of the program to gray-value ranges is necessary:
• The default mode is the adaptive mode, which checks the gray values of an image and automatically adapts
these values for each image. This mode is useful if the gray-value range is unknown or differs between
images. The downside to this default setting is, that the highest value might differ from image to image
which results in the fact that also the curves in the graphs that define the edges appear shifted which can be
confusing.
• The increasing mode checks the values for the first image and keeps these values if the gray-value range
of the following images is either the same or smaller. It only corrects the values for a wider range which is
only a problem if values should be adapted for an image with data that is significantly smaller than the one
of the previous images and the values on the graph are therefore so close together that it is impossible to
distinguish the edges.
• The simplest mode is the fixed mode as it uses only the gray-value range that has been entered and does not
adapt any values. This mode is a good choice if it is known that all images have the same gray-value range
or deviations do not contain necessary information in those values that exceed the given gray-value range so
that their variation of range can be disregarded. You can check your camera to see if 10, 12 or 14 bits are
used and choose your values accordingly.
Depending on the mode you have chosen, the minimum and maximum values will differ. The gray-value range
that is chosen here directly affects the Min. Edge Amplitude as well the graphs on the Fuzzy tab.
The Reset button allows you to set the values back to their default.
If you are not finished with your Input (page 220) yet, proceed to either add calibration images or perform a
calibration, respectively (page 223), if you want to transform your results in to world coordinates, or continue to
create an ROI (page 220) and then step to the next tab to extract edges (page 221).
7.4.5 Advanced Measuring Tasks 225
Assistants
So far, an edge amplitude was used for choosing edges. This is, however, sometimes not sufficient. When, for
example, reflections are part of the image, it might be necessary to further specify the features of the edges that
should be detected. Such features, like position, contrast, pair width or mean gray value can be selected and graded
by using fuzzy measuring.
Fuzzy measuring is based on fuzzy logic and allows a more specific determination of edge selection by assigning
a certain score to each edge that determines whether this edge is a member of a particular fuzzy set. For most
applications, however, it is, not necessary to use fuzzy measuring because the general edge detection functions are
sufficient to detect the right edges. If you want to learn more about fuzzy measuring, please refer to the Solution
Guide III on 1D Measuring. All different fuzzy features are explained in the section “Features that Can Be Used
to Control the Selection of Edges and Edge Pairs”.
In order to use the fuzzy measuring function of the Measure Assistant, you first have to enable it by activating the
Use Fuzzy Measure (Advanced) checkbox on top of the Fuzzy tab.
You can then proceed to select the following fuzzy membership criteria:
These options apply to both edges and edge pairs:
When working with edge pairs, additionally the following criteria can be activated:
• Fuzzy Pair Center Position chooses edge pairs with a center of a certain position,
• Fuzzy Pair Width chooses pairs of a certain width,
• Fuzzy Pair Gray Mean (page 227), selects pairs of a certain mean gray value.
Fuzzy Threshold and Reference Pair Width are two general settings that can be specified and will then be
applied for all activated criteria. Both settings are grayed out until at least one criterion is activated.
The Fuzzy Threshold is a value between 0.1 and 1 that selects the minimum fuzzy score. Each active fuzzy set,
the values added to a fuzzy membership criterion after enabling it, will be evaluated. The final Fuzzy Score is
the geometric mean of the individual scores.
Reference Pair Width helps you to adapt your values to changes in the setup. More information on how to use
Reference Pair Width and the Normalize function can be found in the section “Advanced Fuzzy Features”
(page 227).
How to specify the values that are to be evaluated as “good” by fuzzy measuring is described in more detail below.
In the graph beneath Tolerance, you can see the values corresponding to all extracted edges. These are
displayed as little crosses and the curve that was defined by your good values and the allowed tolerance is
displayed as well.
When you have determined all relevant edges, continue to the Results tab (page 222).
Fuzzy Contrast
Enable Fuzzy Contrast to choose edges with specific amplitudes. Then specify your “good” values as
described in the section “Specify Good Values” (page 225).
Whenall relevant edges have been found, continue to the Results tab (page 222).
Assistants
a)
c) b)
Figure 7.6: How to use fuzzy measuring to improve the edge detection in an image.
The images of a dip switch show how to use fuzzy measuring to improve edge detection. Due to surface
reflections, one of the edge pairs is not detected properly. It is, however, known that the pair width is
9. Setting Fuzzy Pair Width to 9 results in the exclusion of the wrong edge and therefore also in the
detection of all the right edge pairs.
Please note the description about how to specify “good” values (page 225).
You can learn how to use the Normalized option in the paragraph “Advanced Fuzzy Features”.
Whenall relevant edges have been found, continue to the Results tab (page 222).
7.4.5.3 Alignment
If the object in your image appears shifted or rotated, an alignment, which can be chosen on the Code Generation
tab under Alignment Method . Affine Transformation, is necessary. An alignment allows you to automat-
ically relocate your ROI in an image where the object occurs in a different position. Note, however, that the
Measure Assistant does not perform the alignment by itself, but only generates code to align the measurements
if a transformation matrix is available from other methods such as Template Matching. For general information,
please read the following paragraph about alignment with the Measure Assistant. For more detailed information,
please refer to the documentation on matching in the Solution Guide on Matching, Chapter 2.4 “Use the Results
of Matching”.
To perform an alignment with the Measure Assistant, please follow the steps that are described in this paragraph.
When generating code with an activated Affine Transformation option, the assistant produces code that re-
quires a transformation matrix. To find the object within the image, you can perform a matching as explained in
the Solution Guide on Matching or use the Matching Assistant to guide you through the matching task. By teach-
ing a template, the object can be found in any position and orientation. After this successful matching, the affine
transformation can be constructed with the operator vector_angle_to_rigid using the difference between the
original and new model positions. This operator produces the homogeneous transformation matrix that describes
the transformation from the old to the new position.
If the object in your image is only shifted but not rotated, the suitable Alignment Method would be Translation
Only which also needs less processing time than Affine Transformation.
No Alignment is needed if the object always appears in the same position, or the intent is to find the object
location by measuring.
You can now continue to change variable names (page 224) if necessary, or preview the generated code (page 224)
before inserting it into your HDevelop program.
The Measure Assistant Reference provides pull-down menus, a tool bar, tabs with dialogs for most of the tasks,
and a status bar at the bottom in which messages are displayed. Please note that the status bar does not provide a
scrolling mechanism. If the displayed message is to long, move the mouse over it, so that a tool tip displaying the
full message pops up. Alternatively, if the message is only slightly larger than the status bar, you can also drag the
left or right border of the Measure Assistant window to enlarge it.
Images are displayed in the graphics window of HDevelop.
Load Image
You can load an image from a file by the menu item File . Load Image or via the corresponding button of
the tool bar. To acquire images from a camera, you can also use the Snap and Live buttons in the tool bar.
Close Dialog
When closing the Measure Assistant dialog with the menu item File . Close Dialog or the X in the top
Assistants
right corner of the window, the current settings are stored for the duration of the current HDevelop session.
As long as you do not exit HDevelop, you can again open the Measure Assistant with the same settings.
In contrast to this, when you exit the Measure Assistant, the settings are lost also for the current HDevelop
session.
Exit Assistant
When you exit the Measure Assistant with the menu item File . Exit Assistant, the assistant’s dialog
is closed and the current settings are lost unless you have stored them via the menu item File . Save
Current Assistant Settings (page 228). If you want to close the dialog but keep its settings for the
current HDevelop session, you should use the menu item Close Dialog instead.
Draw Line
To create a linear ROI, select the menu item Measuring . Draw Line (also accessible as tool bar button).
For more information about how to draw a linear ROI, please refer to the section “Create ROI” (page 220).
You can also check the ROI data via the tool bar button View ROI Data read more about ROI data in the
section Create ROI (page 220).
Via the menu Help you can access the online documentation.
Assistants
The Code Generation tab includes the following subdivisions:
The OCR Assistant of HDevelop is a front-end to HALCON’s optical character recognition. Using the OCR
Assistant you can, for example,
• easily and quickly set parameters for optical character recognition (OCR) with the Quick Setup,
• segment text by choosing parameters suitable for the characters appearance (page 234)
• choose a pretrained classifier or train your own classifier (page 236),
• choose the kind of results you need (page 240), and
• generate code for your OCR application (page 241).
Note, the assistant is currently limited to OCR presented in the Solution Guide I, chapter ’OCR’ and does not
include Deep OCR.
Using the OCR Assistant is simple: either choose the Quick Setup to load an image and perform an OCR by
setting basic parameters or use the sections Image Source (page 233) and Region of Interest (page 233)
where you can also load a sample and mark the text that should be read with a rectangle. Then improve the
segmentation (page 234) by adapting relevant parameters, choose a pretrained font or performing your own training
(page 236) and finally add the resulting code to your application (page 241).
When looking for an overview of all OCR Assistant elements, please refer to the reference (page 242). The general
process of an OCR application is visualized in figure 7.7. This figure shows how a sample is found in an image
via segmentation and can be directly classified if an OCR classifier is available. The sample is then assigned to a
certain class, a symbol. If no suitable classifier is available, samples can be added to a training file from which a
classifier can be trained that can subsequently be used to classify a sample. The symbol class is typically equivalent
to a simple character. Therefore, a sample that is assigned to a symbol class results in a certain character that is
read.
Segmentation Classification A A
Image Sample Symbol Character
Teaching
Train
Training File Classifier
Optical Character Recognition (OCR) Optical Character Recognition is the technical term for reading and
identifying symbols. In HALCON, OCR is defined as the task to assign an interpretation to regions of
an image. These regions typically represent single characters and therefore we consider this as reading
single symbols.
Sample A sample is the smallest individual object that is detected during segmentation and typically represents a
simple character. It belongs to a certain class, a symbol.
Segmentation Both for the training and for the online reading process, samples must be extracted from the image.
This step is called segmentation. This means that the OCR operators like do_ocr_single_class_svm do
not search for the characters within a given region of interest, but expect a segmented region, which then
will be assigned to a symbol class with a classifier.
Training The training consists of two important steps: First, for each character a number of samples is selected
and stored in training files. In the second step, these files are input to create a new OCR classifier. HAL-
CON provides pretrained OCR classifiers, that are ready-to-use classifiers, which already solve many OCR
applications. These OCR classifiers can be found in the subdirectory ’ocr’ of the directory where you have
installed HALCON.
Font A font describes a certain typeset, that differs from other fonts by certain features of the characters. A
classifier can be trained for these special features of the font so that characters belonging to this font can be
read successfully.
OCR Classifier An OCR classifier is trained to classify a certain set of characters, defined by certain characteris-
tics (for example Pharma.omc). HALCON provides you with a set of pretrained OCR classifiers, which are
based on a large amount of training data from various application areas. These OCR classifiers for fonts al-
low you to read text in documents, on pharmaceutical or industrial products, dot prints, and even handwritten
numbers. Furthermore, HALCON includes pretrained OCR classifiers for fonts like OCR-A and OCR-B.
You can also use your own classifiers and train an OCR classifier with HDevelop.
Symbol A symbol is a class one or more samples are assigned to with the help of an OCR classifier.
Character A character can be a single letter, number or special sign, like a hyphen. It belongs to a certain Font.
7.5.2 Setup
The OCR Assistant allows you to quickly and easily set up your OCR task in just a few steps.
The Quick Setup is an easy way of setting up an OCR application. However, this preconfiguration of segmenta-
tion parameters is based on a basic description of the image content that does not take all factors into consideration
that may influence a segmentation. For this reason, the Quick Setup alone may not be sufficient to segment and
read characters in an image. Therefore, it is recommended and usually necessary to improve an OCR application
by adapting parameters (page 234).
To get started, it is very easy to set up an OCR task with the OCR Assistant’s Quick Setup in five steps:
1. Load a sample image by clicking the icon right beside the text or alternatively acquire a live image by
selecting the Image Acquisition Assistant (page 180) under Image Source on the same tab.
2. Mark the position of the text to be read using a rectangle. The purpose of this step is to indicate the text
geometry (like size and rotation). Use an axis-aligned rectangle for text in horizontal or vertical direction
and a rotated rectangle for rotated text. Alternatively, set the rectangle via the Region of Interest dialog
on the same tab. Allow for a small border. Note that if you are using the Quick Setup the rectangle
must enclose the characters that should be read as precisely as possible and it has to include at least three
characters.
3. Enter the text that you expect to be read in the text field. Remember to enter the text exactly as it is in the
sample, including all separators, line breaks and correct capitalization. If the text contains large gaps these
should be mimicked for a better recognition.
7.5.2 Setup 233
4. Describe some basic properties of the text. These properties are essential for a successful segmentation
with a minimum number of parameter adaptations. Without this information, the samples in the image may
Assistants
not be segmented at all (for example, for characters that are printed bright on dark background, the image
has to be inverted internally). Check if characters are bright on dark background, composed of individual
dots or if the text is structured, that is if letters, digits or separators appear at fixed positions. Furthermore,
check whether the background is textured, noisy or cluttered.
5. Check if all information that was required so far is filled in correctly according to the features of the sample
image and then click the Apply Quick Setup button when finished.
If the Quick Setup was successful, the found characters are read and the Results tab (page 240) is automatically
opened. Otherwise, the segmentation tab opens and allows you to adapt parameters while simultaneously observing
the segmentation results.
The Image Source dialog on the Setup tab lets you choose images from different sources:
• If your image has already been opened in the Graphics Window, you can activate Graphics Window to
continue working with the currently displayed image.
• You can load an image file by activating Image File or choosing Load Image from the menu File and
the toolbar, respectively, and either typing in the image path or using the Browse button on the right to select
an image from a file.
• Another option is to activate the radio button for the Image Acquisition Assistant (page 180). Being con-
nected to this assistant, you can acquire the image you want to use for OCR and you can even choose to
work with a Live image. The same operations can be performed by clicking the Snap or Live symbols in
the toolbar.
Note that the OCR works on a single channel image. For color RGB images, the red channel will be used. A color
transformation can be performed with the operator trans_from_rgb later in the generated code.
When you have loaded your image, you can continue to create an ROI.
1. Create a ROI by using the buttons Draw Axis-aligned Rectangle or Draw Rotated Rectangle in the
tool bar or in the drop-down menu OCR. You can also choose a shape within the Region of Interest
section of the Setup tab to create a ROI. The shape of the ROI should be chosen according to the shape of
the object to be inspected.
2. After having chosen the ROI shape, you “draw” the ROI in the Graphics Window by keeping the left mouse
button pressed. Then, modify the ROI until it has the correct shape. When you are finished, click the right
mouse button, to confirm your choice.
3. To select a rotated rectangle with a correct reading direction, check that the arrow that is displayed within
the rectangle corresponds to the reading direction of the characters.
All characters that can be read with default parameters within the ROI are displayed immediately.
Note that the ROI should be chosen such that it also contains a small border around the characters to be recognized.
If you have previously prepared and saved an ROI, you can reuse it by selecting Load ROI from the OCR menu.
In order to delete one or more ROIs, you can either mark them and just press the delete button on your keybord or
click the buttons for Delete Selected ROI Item or Delete All ROIs on the Setup tab.
You can view and edit the ROI data, which includes all the data about the exact position of the ROI(s), using the
tool bar button View ROI Data. Editing an ROI by changing the ROI data is useful if an ROI should be modified
more precisely than is possible by drawing in the graphics window.
When you are satisfied with the shape of your ROI, proceed to adjust parameters on the Segmentation tab.
234 HDevelop Assistants
7.5.3 Segmentation
This section describes how to set segmentation parameters for an OCR application. The tab Segmentation al-
lows you to adjust several kinds of parameters for optical character recognition. The success of these parameter
adaptions can be viewed immediately in the image after the Setup (page 232) has been completed, if at least one
image has been loaded and an ROI has been specified. Adjusting parameter settings is also recommended when
using the Quick Setup (page 232) to achieve the highest possible character recognition quality.
Here three options specifying the way the characters are printed, can be chosen.
• Light-On-Dark whenever the gray values of the printed characters are brighter than the background gray
values,
• Dot Print if the character print is composed of single dots, for example, from matrix printers, and
• Imprinted / Embossed for print that is visible more by its relief rather than its pigment.
This dialog allows you to choose up to three options that specify the shape of a symbol.
• Interpunctuation for a text containing punctuation like, for example, comma or period,
• Separators for a text containing separators such as hyphens or slashes, and
• ALL UPPER CASE for a text containing numbers and upper case letters only.
If the orientation of the text is unknown, for example when it appears in variable orientations during the application,
activate Line Orientation. Select an angle range by adjusting the minimum and maximum angle in degrees to
automatically correct for text that is rotated within a given range. Note that not only the orientation but also the
reading direction is also relevant for setting a value for Text Orientation. This is especially important for
vertical text.
If text may appear slanted, activate Symbol Slant to automatically correct for text that is printed in slant or italics.
Each value can be reset to the default value by clicking the corresponding buttons on the right side.
7.5.3 Segmentation 235
Assistants
This dialog offers several options to make adjustments specifying a certain layout for text to be read. Those
adjustments apply to characteristics of the text lines as they appear within the image.
• Baseline Tolerance, sets the allowed amount of deviation from a perfect line, via text field or slider.
• Max. sets the Number of Lines, and therefore limit the number of text lines that will be searched for, via
text field or slider.
• Line Structure sets the maximum number of words and characters when searching for lines, for example
’3 4-5’ indicates a line consisting of three letters, then a space, then 4-5 letters.
• Activate the checkbox of Eliminate Background Lines if background lines exist, for example, in certain
forms or scanned documents.
The values of Baseline Tolerance and Max. Number of Lines can easily be reset to the default values via
the buttons on the right side of the sliders.
7.5.3.7 Inspection
The Inspection section helps you to find a solution to a segmentation problem. Therefore, the embedded window
shows only the content of the region of interest. By choosing an option from the combo box, intermediate results
for different parameters as well as filtered images or symbol candidates can be viewed. By adapting parameters
on this tab, changes in the segmentation process may be observed within this window which may be more helpful
than just viewing the final result in the Graphics Window.
The following table can help you to solve segmentation problems by adapting the correct parameters.
Symbol Results No results or incorrect results are found. Check if the parameters in the dialogs Symbol
Appearance, Size and Shape correspond to the
appearance of the characters, that is, if the char-
acters are composed of single dots (dot print), the
corresponding parameter has to be set.
Please note that if there are problems with the OCR, first of all the image quality should be checked. Only if there
is nothing to enhance about the quality and therefore the training images are representative for the images in the
application, parameters should be changed.
7.5.3.8 Reset
To simply set everything in this tab back to the default settings, which is recommended if a new OCR application
is started with the OCR Assistant, just press the Reset All button. To reset single values on this tab, use the
corresponding buttons on the right side of the selections.
This section deals with the classifier that is used for the OCR application and with its settings.
When using a previously trained classifier, you can choose between either
• a Pretrained Classifier from one of the classifiers that are available for HALCON (see Solution Guide
I, chapter ’OCR’ for a more detailed description of the available pretrained OCR classifiers) or
• a previously trained classifier from file that can be selected via the Browse button next to the combo box for
HALCON classifiers.
When deciding which pretrained classifier should be used, it may help to view details of the OCR classifier by
clicking the button with the magnifying lens located right of the selection for pretrained classifiers.
If a new classifier should be trained, a Training File has to be used as a basis for training a new classifier.
Therefore, select the radio button Training File and continue by loading a training file with the Browse button
next to the text field. To start the creation of a new training file, click the button New. After a new training file
has been created, samples can be added to the training file and then it has to be saved via the button Save before
it can be used for training. To look at a currently or previously created training file and maybe edit it, open the
Training File Browser (page 86) which provides an overview of the content of training files which can easily
be modified.
7.5.4.2 Teaching
This section allows you to ’teach’ characters which are displayed in the window right next to the text field.
Steps for teaching:
1. Use the text field to enter characters that correspond to the segmented characters in the image. As you
enter the characters into the text field, the image of the corresponding character is displayed on the right side
of the text field. The characters are highlighted in the Graphics Window as well.
2. Click the button Add To Training Data.
3. Repeat this process for each sample image.
5. When the classifier has been trained for the first time, results of the OCR are available in the text field at
the bottom of the Teaching dialog to assist with further teaching.
Assistants
6. If the results of the OCR are correct, click the button Suggestion to quickly add the results to the text
field during further teaching.
If you want to view the results of the OCR classification, continue to the Results tab (page 240).
If you want to improve the OCR classification, you can change Basic Features that may improve a classification
and train again. You can also select a different classifier and change parameters that influence single classifiers via
the dialog Advanced Training Parameters and Features.
If you want to save data for archival purposes during training, you can activate Save Ground Truth. If activated,
each time data is added to the training file, a copy of the source image, assistant settings, and teach data will be
stored in a subdirectory next to the training file.
7.5.4.3 Training
The Training Dialog gives you the opportunity to perform a training via the button Train Now and is connected
to the more interactive Teaching dialog (page 236) above.
Other than starting the training, the status of the training can be viewed including
• the Number of Samples that were used in the selected training file,
• the name and path of the Classifier,
• and the status of the classifier, ’untrained’, ’training’, ’trained’, ’outdated’ (new training data is available that
has not been trained yet), ’read only’ (classifier has been loaded) and ’failed’ (the last training has not been
successful).
This dialog enables you to choose several basic features that affect the classification.
Interpolation
The parameter Interpolation lets you choose the interpolation mode, that is the adaptation of characters
in the image to the pattern size. It also influences the parameter Gray Values as this, if activated, uses the
interpolated pattern. The most suitable interpolation method largely depends on the values that were chosen
as Pattern Width and Pattern Height, as the scale factors.
Weighted Bilinear interpolation in an image with Recommended choice if characters are scaled down
a Gaussian filter. by a large amount and a very high precision is re-
quired. Note that this interpolation method requires
more processing time.
Bilinear Interpolation, using the values of the 4 Can be used if characters are not scaled very much.
closest pixels in diagonal direction.
Nearest Neigh- No interpolation is performed. Fast but not very precise interpolation. Should only
bor be used if the image is blurred.
This selection allows a precise specification of the classifier as well as of the classification parameters.
The following table can help you choose suitable classifier parameters.
Components MLP, Number of components (de- Dimension of the reduced feature space when using ap-
SVM fault = 10) propriate preprocessing. For preprocessing with prin-
Assistants
cipal components or canonical variates the length of
the data is determined in Components. It is ignored if
Processing was set to None or Normalization.
Hidden Units MLP Number of hidden units Number of neurons hidden in the middle layer of the
(1...150, default = Auto) MLP. The more input data you are using, the higher
this value should be. In many cases, very small values
of Hidden Units already lead to very good classifica-
tion results. If Hidden Units is chosen too large, the
MLP learns the training data very well, but does not re-
turn very good results for unknown data. In the ’auto’
mode, the hidden value is estimated by a heuristic al-
gorithm which is based on the number of characters. If
the number of characters is really high, the number of
hidden units might be estimated too high, which leads
to a very slow training.
Iterations MLP The maximum number of it- Select a sufficient number of iterations to create an
erations for training a MLP MLP classifier that performs well in the subsequent ap-
classifier (default = 200) plication.
Weight Toler- MLP MLP training continues Choose a realistic value for Weight Tolerance. If
ance while weights still change this value is chosen too small in the training, the train-
more than this value be- ing will take longer. If Weight Tolerance is set very
tween iterations (default = high, the training will abort quickly and the classifica-
1) tion results will not be very useful either.
Error Toler- MLP MLP training continues Choose a realistic value for Error Tolerance. If this
ance while error still changes value is chosen too small in the training, the training
more than this value be- will take longer. If Error Tolerance is set very high,
tween iterations (default = everything is accepted which means that the results
0.01) may not be useful either.
Mode SVM One vs. All, One vs. One The voting method used to combine the binary sup-
(default = One vs. All) port vector machine classifiers. One vs. All creates
a classifier where each class is compared to the rest
of the training data. During testing, the class with the
largest output is chosen. One vs. One creates a bi-
nary classifier between each single class. During test-
ing a vote is cast and the class with the majority of the
votes is selected.
Specialization SVM The Gamma parameter of It specifies the amount of influence of a support vector
(Gamma) the radial basis kernel func- upon its surroundings. A big value means a small in-
tion. 0.01, 0.02, 0.05, 0.1, fluence of surroundings, each training vector becomes
0.5 (default = 0.02) a support vector and training/classification times grow.
A too small value leads to few support vectors.
Regularization SVM Regularization constant of The Nu parameter of the radial basis kernel function.
(Nu) an SVM (default 0.05) One typical strategy is to select a small Specialization
(Gamma) and Regularization (Nu) pair and consecu-
tively increase the values as long as the recognition rate
increases.
Number of kNN Number of trees in a tree The number of trees used by the kNN classifier. If more
Trees structure (default = 4) trees are used, the classification becomes more robust
but the runtime also increases.
Other features that can be chosen are pixel, pixel_invar, pixel_binary, gradient_8dir,
240 HDevelop Assistants
7.5.5 Results
The Results tab lets you select features concerning the classification results.
When you are satisfied with the results, proceed to the Code Generation tab to receive the code for your appli-
cation. Otherwise, go back to adapt your segmentation (page 234) or OCR classifier parameters (page 236).
To use word processing, activate Enable on the right side of this dialog. To correct the result of the OCR classifi-
cation according to certain rules, the assistants’ word processing offers two options:
Note that, in contrast to the operators do_ocr_word_*, the entire word is used in the Regular Expression
automatically, that is the given expression is surrounded by ’ˆ...$’.
Furthermore, the maximum number of characters that will be corrected can be selected via Max. Corrections.
For up to this number of characters if the word does not match the expression one or more alternatives, like the
second best class and so on are considered, depending on the number set for Symbol Alternatives.
Note that if high numbers of Max. Corrections and Symbol Alternatives are chosen the number of possible
corrections may be very high. Therefore, if the number of characters to be read is very high, the number of allowed
corrections is internally clipped to 5, 3, or 1 if the alternatives multiplied with the number of characters are equal
or greater than 30, 60 or 90, respectively.
The necessary corrections also influence the Score (the more corrections the lower the score) of the result that can
be viewed in the Results tab.
For more information on how to use the word processing parameters, please refer to the reference documentation
of the operators do_ocr_word_*.
You can select one or more of the following features, depending on whether you have enabled Word Processing
or not. The results for this feature are then displayed in the Results section.
Without word processing, you can enable
• Symbol to view the class determined for the symbol,
• Confidence to see with which confidence the class of the symbol was determined, and
• Alternatives to view which classes best match the symbol.
With word processing enabled, you can select
• Word to display the text that was read, including corrections,
• Uncorrected Word to see the text that was read without corrections, and
• Score to see how well the uncorrected text matched the expected format on a scale from 0 to 1.0 (where 0 is
failure to correct and 1 is a match without corrections).
7.5.6 Code Generation 241
Assistants
To adapt the visualization, several display parameters can be adjusted.
The two combo boxes Symbol Color and Text Color let you select colors for text visualization. Via Symbol
Color, a color, or color scheme that will be used to mark the extracted characters, can be selected. Text Color
on the other hand lets you choose the color of the text that is read.
Use the dialog Display Font that opens when clicking the font button to choose the font as well as further font
settings for the text that is read
Several display functions can be turned on and off. If activated,
7.5.5.4 Results
The Results section consists of two windows. The first one shows the active ROI and also lets you choose
between the different ROIs so you can view their results. The results of each ROI are displayed in the text field
below. Columns will be displayed according to the features chosen in Feature Selection (page 240).
Code Generation produces the code that is necessary to perform the chosen OCR tasks within an HDevelop
program. On the Code Generation tab you can choose between several options and change parameter names,
which has a direct effect on the code that is generated.
Under General Options, you can decide on the Generation Mode, on the Alignment Method and whether
code from the Image Acquisition Assistant should be integrated into the generated code.
Generation Mode lets you select the task for which the code is generated, that is
• Text Reading or
• OCR Classifier Training.
Activate Initialize Acquisition if you have previously used the Image Acquisition Assistant (page 180) to
acquire images and want to integrate the generated code from this assistant in your code. Such an automatic inte-
gration of code results, for example, in the automatic opening of the framegrabber if Image Acquisition Assistant is
activated. If code for image acquisition already exist, the checkboxes should be deactivated. If Image Acquisition
Assistant is not in use, Initialize Acquisition is grayed out.
Alignment
If the characters in your image can appear shifted or rotated, an alignment, which can be chosen on the
Code Generation tab under Alignment Method, is necessary. An alignment allows you to automatically
relocate your ROI in an image where the object occurs in a different position. Note, however, that the OCR
Assistant does not perform the alignment by itself, but only generates code to align the measurements if a
transformation matrix is available from other methods such as Template Matching. For general information,
please read the following paragraph about alignment with the OCR Assistant. For more detailed information,
please refer to the documentation on matching in the Solution Guide on Matching, Chapter 2.4 “Use the
Results of Matching”.
To perform an alignment with the OCR Assistant, please follow the description in this paragraph. When gen-
erating code with an activated Affine Transformation option, the assistant produces code that requires a
242 HDevelop Assistants
transformation matrix. To find the characters within the image, you can perform a matching as explained in
the Solution Guide on Matching or use the Matching Assistant (page 202) to guide you through the matching
task and generate the alignment code. No Alignment is needed if the characters always appear in the same
position. You can now continue to change variable names if necessary, or preview the generated code before
inserting it into your HDevelop program.
If desired, you can change the default variable names or replace them with your own variable names.
Before clicking the Insert button to include your code into the program window, you can preview the code in the
Code Preview table. You can now step through the table which consists of the columns
• Insert Operator, which shows the operator that will be inserted when you press the Insert button,
• Procedure, which shows the corresponding procedure,
• Line, referring to the line number within the code, and
• Replace Operator, which shows previously generated code that will be replaced.
• When you are finished changing parameters, click the Insert Code button on the right side of the General
Options dialog to generate the code.
• Finally, integrate the code into your HDevelop program.
The OCR Assistant provides pull-down menus. Tabs with the dialogs for most of the tasks that can be performed
with the Measure Assistant, and a status bar at the bottom in which messages are displayed. Please note that the
status bar does not provide a scrolling mechanism; if the displayed message is to long, move the mouse over it,
so that a tool tip displaying the full message pops up. Alternatively, if the message is only slightly larger than the
status bar, you can also drag the left or right border of the OCR Assistant window to enlarge it.
Images are displayed in the graphics window of HDevelop.
Load Image
You can load an image from a file by the menu item File . Load Image or via the corresponding button of
the tool bar. To acquire images from a camera, you can also use the Snap and Live buttons in the tool bar.
Assistants
of the tool bar. Then you can load the training file again in a later session.
Via the menu Help you can access the online documentation.
Assistants
The Results tab consists of the following subdivisions:
Chapter 8
Language
HDevelop Language
This chapter introduces the syntax and the semantics of the HDevelop language. It illustrates what can be entered
into a parameter slot of an operator or procedure call. In the simplest case this is the name of a variable, but it
might also be an arbitrary expression like sqrt(A). Besides, control structures (like loops) and the semantics of
parameter passing are described.
Note that the HALCON operators themselves are not described in this chapter. For this purpose refer to the
HALCON reference manual. All program examples used in this chapter can also be found in the directory
%HALCONEXAMPLES%\hdevelop\Manuals\HDevelop.
HALCON distinguishes two kinds of data: control data (numbers, strings, or handles) and iconic data (images,
regions, etc.) By further distinguishing input from output parameters, we get four different kinds of parameters.
These four kinds always appear in the same order in the HDevelop parameter list. In the reference manual operator
signatures are visualized in the following way:
Iconic input objects are always passed first, followed by the iconic output objects. The iconic data is followed by
the control data, and again, the input parameters succeed the output parameters.
Any of the four types of parameters may be empty. For example, the signature of read_image reads
The operator read_image has one output parameter for iconic objects Image and one input control parameter
FileName. The parameter types are reflected when entering operators in the operator window. The actual operator
call displayed in the HDevelop program window is:
The parameters are separated by commas. Input control parameters can either be variables, constants or expres-
sions. An expression is evaluated before it is passed to a parameter that receives the result of the evaluation. Iconic
parameters must be variables. Control output parameters must be variables, too, as they store the results of an
operator evaluation.
248 HDevelop Language
All non-iconic data is represented by so called control data (numbers, strings or handles) in HDevelop. The name
is derived from their respective functions within HALCON operators where they control the behavior of image
processing, for example, thresholds for a segmentation operator. Control parameters in HDevelop may contain
arithmetic or logical operations. A control data item can be of one of the following data types: integer, real,
string, boolean, and handle.
integer The data type integer is used under the same syntactical rules as in C. Integer numbers can be input
in the standard decimal notation, in hexadecimal by prefixing the number with 0x, and in octal by prefixing
the number with 0 (zero).
For example:
4711
-123
0xbeef (48879 in decimal notation)
073421 (30481 in decimal notation)
Data items of type integer are converted to their machine-internal representations, that is the C type long
(4 or 8 bytes).
The minimum and maximum possible values of an integer depend on whether the 64- or 32-bit version of
HALCON is used. You can use the operator tuple_number to verify that a value fits into the range of the
integer data type. For a reference, see table 8.1.
64-bit 32-bit
Minimum value −9223372036854775808 −2147483648
Maximum value 9223372036854775807 2147483647
real The data type real is used under the same syntactical rules as in C.
For example:
73.815
0.32214
.56
-17.32e-122
32E19
Data items of type real are converted to their machine-internal representations, that is the C type double
(8 bytes).
string A string is a sequence of characters that is enclosed in single quotes ('). Special characters, like the line
feed, are represented in the C-like notation, as you can see in table 8.2, see the reference of the C language for
comparison. You can enter arbitrary characters using the format \xnn where nn is a two-digit hexadecimal
number, or using the format \0nnn where nnn is a three-digit octal number. Less digits may be used if the
string is unambiguous. For example, a line feed may be specified as \xa unless the string continues with
another hexadecimal digit (0-F).
For example: The string Sobel's edge-filter has to be specified as 'Sobel\'s edge-filter'. A Win-
dows directory path can be entered as 'C:\\Programs\\MVTec\\Halcon\\images'
boolean The constants true and false belong to the data type boolean. The value true is internally repre-
sented by the number 1 and the value false by 0. This means, that in the expression Val := true the
effective value of Val is set to 1. In general, every integer value other than 0 means true. Note that some
HALCON operators take logical values for input, for example, set_system. In this case, the HALCON
operators expect string constants like 'true' or 'false' rather than the boolean values true or false.
8.2 Control Types and Constants 249
Language
bell BEL \a
backslash \ \\
single quote ' \'
arbitrary character (hexadecimal) \xnn
arbitrary character (octal) \0nnn
handle Handles are references to complex data structures, for example, a connection to an image acquisition
device or a model for shape-based matching.
In addition to these general types, there are special constants and the type tuple, which are specific to HALCON
or HDevelop, respectively. HDevelop also supports the variable type vector, see section 8.6 on page 270.
constants For the return value (result state) of an operator, constants exist. The constants can be used together with
the operator dev_error_var and dev_set_check. These constants represent the normal return value of
an operator, so-called messages. For errors, no constants are available but there are plenty of error numbers
internally, see the Extension Package Programmer’s Manual, appendix A on page 105.
In table 8.3, all return messages can be found.
Additionally, there are constants for the types of control data, see table 8.4. These can be compared to the
result of a type operation to react to different types of control data, see section 8.5.5 on page 259.
Finally, you can access minimum and maximum values of integers and floating-point numbers via the con-
stants mentioned in table 8.5.
250 HDevelop Language
Constant Meaning
H_FLOAT32_EPSILON maximal relative approximation error for 32-bit floating-point numbers
H_FLOAT32_MAX largest possible 32-bit floating-point number
H_FLOAT32_MIN smallest possible 32-bit floating-point number
H_FLOAT32_MIN_POSITIVE smallest possible positive floating-point number (32-bit)
H_FLOAT64_EPSILON maximal relative approximation error for 64-bit floating-point numbers
H_FLOAT64_MAX largest possible 64-bit floating-point number
H_FLOAT64_MIN smallest possible 64-bit floating-point number
H_FLOAT64_MIN_POSITIVE smallest possible positive 64-bit floating-point number
H_FLOAT_INFINITY ∞ (for example: 1.0/0.0), shown as inf in variable window
H_FLOAT_NAN Not a Number (NaN), shown as nan in variable window
H_FLOAT_NEG_INFINITY −∞ (for example: −1.0/0.0), shown as -inf in variable window
H_INT32_MAX largest possible 32-bit integer
H_INT32_MIN smallest possible 32-bit integer
H_INT64_MAX largest possible 64-bit integer (error in 32-bit HALCON)
H_INT64_MIN smallest possible 64-bit integer (error in 32-bit HALCON)
H_INT_MAX H_INT64_MAX or H_INT32_MAX, depending on HALCON version
H_INT_MIN H_INT64_MIN or H_INT32_MIN, depending on HALCON version
For further explanation of the numeric constants, see the HDevelop example program
tuple_numeric_limits.hdev. See also the description of the operator tuple_constant.
tuple The control types are only used within the generic HDevelop type tuple. A tuple of length 1 is interpreted as
an atomic value. A tuple may consist of several data items with different types. The standard representation
of a tuple is a listing of its elements included into brackets. This is illustrated in figure 8.1.
Value string
integer
Value
[ ]
real
boolean
,
handle
[] specifies the empty tuple. A tuple with just one element is to be considered as a special case, because
it can either be specified in the tuple notation or as an atomic value: [55] defines the same constant as 55.
Examples for tuples are:
[]
4711
0.815
8.3 Variables 251
'Text'
[16]
[100.0,100.0,200.0,200.0]
['FileName','Extension']
[4711,0.815,'Hugo']
8.3 Variables
Language
The names of variables are built up by composing letters, digits, and the underscore ‘_’. The type of a variable,
iconic or control variable, depends on its position in the parameter list in which the variable identifier is used for
the first time, see also section 8.1 on page 247, and is determined during the input of the operator parameters.
Whenever a new identifier appears, a new variable with the same identifier is created. Control and iconic variables
must have different names. The value of a variable is undefined until the first assignment defines it, note that the
variable has not been instantiated yet. A read access to an undefined variable leads to a runtime error.
Instantiated variables contain tuples of values. Depending on the kind of the variable, the data items are either
iconic objects or control data. The length of the tuple is determined dynamically by the performed operation. A
variable can get new values any number of times, but once a value has been assigned the variable will always keep
being instantiated, unless you select the menu item Menu Execute . Reset Program Execution. The content
of the variable is deleted before the variable is assigned new values.
The concept of different types of variables allows a first (“coarse”) typification of variables (control or iconic data),
whereas the actual type of the data for example, real, integer, string etc., is undefined until a concrete value
is assigned to the variable. Therefore, it is possible that the data type of a new data item differs from that of the
old.
In HDevelop, the type of a variable is defined in three different ways:
In the latter case, the type of a variable is defined by those code lines, where the value of the variable is used and
where the exact type of the new value is known before run time.
Within its scope, a variable must always be properly defined. The following type definition errors are possible:
• A local variable is used, whose type has not been defined before.
→ The error ’The variable never gets initialized’ [21089] is thrown. All lines using the variable will become
invalid, and the variable will not be displayed in the variable window.
• The type of a variable is defined differently, for example, as iconic variable and later as control variable.
→ The error ’The type of the variable could not be determined (conflicting type definitions found)’ [21088]
is thrown. All lines using the variable will become invalid, and the variable will not be displayed in the
variable window.
• The type of a variable is properly defined, but used in the wrong context, for example, an iconic variable is
used as a control input parameter.
→ The error ’The type of the variable could not be determined (conflicting type definitions found)’ [21088]
is thrown. Only the affected code will become invalid.
252 HDevelop Language
HDevelop supports local and global variables. All variables are local by default. They exist only within their
procedure. Local variables with the same name may exist in different procedures without interfering with each
other. In contrast, global variables may be accessed in the entire program. They have to be declared explicitly
using the global statement.
When using HDevEngine, a global variable can extend beyond the scope of its program as HDevEngine manages
the variables of all loaded programs and procedures. This can be used to transfer information between procedures.
Each global variable can have only one value at a time for all procedure calls of the running HDevEngine instance.
Therefore, take care not to overwrite the value of a variable of one program with that of another.
For more information about HDevEngine, see the Programmer’s Guide, chapter 21 on page 139.
The declaration
main procedure:
* declare global variables
global tuple File
global object Image
...
File := 'particle'
read_image(Image, File)
process_image()
* Image has been changed by process_image()
* File remains unchanged
...
process_image procedure:
* use global variable
global object Image
...
auto_threshold (Image, Regions, 2)
File := 'fuse'
read_image(Image, File)
return()
Procedures have to explicitly announce their use of global variables, existing procedures cannot be broken by
introducing global variables in other parts of the program. By nature, the names of global variables have to be
unique in the entire HDevelop program. The variable window provides a special tab to list all global variables that
are currently declared.
8.4 Operations on Iconic Objects 253
Symbol Types
i integer
a arithmetic, that is: integer or real
b boolean
s string
v all types (atomic)
t all types (tuple)
Language
Table 8.6: Symbols for the operation description.
Please note that in all following tables variables and constants have been substituted by letters which indicate
allowed data types. These letters provide information about possible limitations of the areas of definition. The
letters and their meaning are listed in table 8.6. Operations on these symbols can only be applied to parameters of
the indicated type or to expressions that return a result of the indicated type.
The symbol names i, a, b, and s can denote atomic tuples (tuples of length 1) as well as tuples with arbitrary
length.
Operations are normally described assuming atomic tuples. If the tuple contains more than one element, most
operators work as follows:
• If one of the tuples is of length 1, all elements of the other tuples are combined with that single value for the
chosen operation.
Input Result
5 * 5 25
[5] * [5] 25
[1,2,3] * 2 [2,4,6]
[1,2,3] * 2.1 + 10 [12.1,14.2,16.3]
[1,2,3] * [1,2,3] [1,4,9]
[1,2,3] * [1,2] runtime error
'Text1' + 'Text2' 'Text1Text2'
17 + '3' '173'
'Text ' + 3.1 * 2 'Text 6.2'
3.1 * (2 + ' Text') runtime error
3.1 + 2 + ' Text' '5.1 Text'
3.1 + (2 + ' Text') '3.12 Text'
Table 8.7: Examples for arithmetic operations with tuples and strings.
• If both tuples have a length greater than 1, both tuples must have the same length, otherwise a runtime error
occurs. In this case, the selected operation is applied to all elements with the same index. The length of the
resulting tuples is identical to the length of the input tuples.
• If one of the tuples is of length 0 ([]), a runtime error occurs.
In table 8.7 you can find some examples for arithmetic operations with tuples. Pay special attention to the order in
which the string concatenations are performed. The basic arithmetic operations in HDevelop are +, -, *, /. Please
note that + is a dimorphic operation: If both operands are numeric, it adds numbers. If at least one of the operands
is a string, it concatenates both operands as strings.
8.5.2 Assignment
In HDevelop, an assignment is treated like an operator. To use an assignment, you can either type it directly
into the program window or use the operator assign (Input, Result). Should you choose to work with the
operator window, the operator has the following semantics: It evaluates Input (right side of assignment) and stores
it in Result (left side of assignment). However, in the program text the assignment is represented by the usual
syntax of the assignment operator: Result := Input. The following example outlines the difference between an
assignment in C syntax and its transformed version in HDevelop:
The assignment in C syntax:
u = sin(x) + cos(y);
u := sin(x) + cos(y)
If the result of the expression does not need to be stored in a variable, the expression can directly be used as input
value for any operator, or be typed into the program window. Therefore, an assignment is necessary only if the
value has to be used several times or if the variable has to be initialized, for example for a loop.
Modifying tuple elements:
Inside the program window enter:
8.5.2 Assignment 255
Result[Index] := Value
When using the operator window enter the operator assign_at and fill in the fields.
Language
As an example:
Areas := [1,2,3]
Areas[1] := 9
Tuple := []
for i := 0 to 5 by 1
Tuple[i] := sqrt(real(i))
endfor
As you can see from the examples, the indices of a tuple start at 0.
An insertion into a tuple can generally be performed in one of the following ways:
1. In case of appending the value at the ‘back’ or at the ‘front’, the tuple concatenation operation , (comma)
can be used. Here the operator assign is used with the following parameters:
assign ([Tuple,NewVal],Tuple)
which is displayed as
Tuple := [Tuple,NewVal]
2. If the index position is somewhere in between, the operator tuple_insert has to be used.
To insert the tuple [11,12,13] into the tuple [1,2,3] at position 1, use
or as in-line operation
resulting in [1,11,12,13,2,3].
In the following example, regions are dilated with a circle mask and afterwards, the areas are stored into the tuple
Areas. In this case, the operator assign_at is used.
Please note that first the variable Areas has to be initialized to avoid a runtime error. In the example, Areas is
initialized with the empty tuple ([]). Instead of assign_at, the operator assign with tuple concatenation could
be used because the element is appended at the end of the tuple:
Areas := [Areas,Area]
A basic tuple operation may be selecting one or more values, combining tuples (concatenation), or getting the
number of elements, see table 8.8 for operations on tuples containing control data.
Table 8.8: Basic operations on tuples (control data) and the corresponding HALCON operators.
The concatenation accepts one or more variables or constants as input. They are all listed between the brackets,
separated by commas. The result again is a tuple. Note the following: [[t]] = [t] = t.
8.5.4 Tuple Creation 257
control iconic
[] gen_empty_obj()
[t1,t2] concat_obj(p1, p2, q)
|t| count_obj(p, num)
t[i] select_obj(p, q, i+1)
t[i1:i2] copy_obj(p, q, i1+1, i2-i1+1)
Table 8.9: Equivalent tuple operations for control and iconic data.
Language
|t| returns the number of elements of a tuple. The indices of elements range from zero to the number of elements
minus one (|t|-1). Therefore, the selection index has to be within this range.1
Tuple := [V1,V2,V3,V4]
for i := 0 to |Tuple|-1 by 1
fwrite_string (FileHandle,Tuple[i]+'\n')
endfor
[1,Var,[64,128]] [1,2,4,8,16,16,32,64,128]
|Var| 6
Var[4] 16
Var[4,5] [16,32]
Var[2:4] [8,16,16]
subset(Var,[0,2,4]) [2,8,16]
select_mask(Var,[1,0,0,1,1,1]) [2,16,16,32]
remove(Var,[2,3]) [2,4,16,32]
find(Var,[8,16]) 2
uniq(Var) [2,4,8,16,32]
Other examples can be found in the program tuple.hdev. The HALCON operators that correspond to the basic
tuple operations are listed in table 8.8 on page 256.
Note that these direct operations cannot be used for iconic tuples. Iconic objects cannot be selected from a tuple
using [] and their number cannot be directly determined using ||. For this purpose, HALCON operators are
offered that carry out the equivalent tasks. In table 8.9 you can see tuple operations that work on control data
(and which are applied via assign or assign_at) and their counterparts that work on iconic data (and which are
independent operators). In the table, the symbol t represents a control tuple, and the symbols p and q represent
iconic tuples.
The simplest way to create a tuple, as mentioned in section 8.2 on page 248, is the use of constants together with
the operator assign (or in case of iconic data one of its equivalents shown in table 8.9):
assign ([],empty_tuple)
assign (4711,one_integer)
assign ([4711,0.815],two_numbers)
1 In contrast, the index of objects for example, select_obj, ranges from 1 to the number of elements.
258 HDevelop Language
empty_tuple := []
one_integer := 4711
two_numbers := [4711,0.815]
This is useful for constant tuples with a fixed (small) length. More general tuples can be created by successive
application of the concatenation or the operator assign_at together with variables, expressions, or constants. If
we want to generate a tuple of length 100, where each element has the value 4711, it might be done like this:
tuple := []
for i := 1 to 100 by 1
tuple := [tuple,4711]
endfor
Because this is not very convenient, a special function called gen_tuple_const is available to construct a tuple
of a given length, where each element has the same value. Using this function, the program from above is reduced
to:
tuple := gen_tuple_const(100,4711)
A fast way to create a sequence of values with a common increment is to use tuple_gen_sequence. For example,
to create a tuple containing the values 1..1000, use
tuple_gen_sequence(1,1000,1,Sequence)
Sequence := [1:1:1000]
If the increment value is one (as in the above example), it is also possible to write:
Sequence := [1:1000]
If we want to construct a tuple with the same length as a given tuple there are two ways to get an easy solution.
The first one is based on gen_tuple_const:
tuple_new := gen_tuple_const(|tuple_old|,4711)
Here we get first a tuple of the same length with every element set to zero. Then, we add the constant to each
element.
In the case of tuples with different values we have to use the loop version to assign the values to each position:
tuple := []
for i := 1 to 100 by 1
tuple := [tuple,i*i]
endfor
In this example, we construct a tuple with the square values of i from 12 to 1002 .
8.5.5 Type Operations 259
Language
b := is_valid_handle(t) test for valid handles tuple_is_valid_handle
s := sem_type(t) get semantic type tuple_sem_type
i := type(t) get type value tuple_type
The type operations allow you to test or query the value type of control data as shown in table 8.10. See table 8.4
on page 249 for the corresponding type constants.
There are also corresponding operations that test each element of the input tuple as shown in table 8.11.
See table 8.12 for an overview of the available basic arithmetic operations.
All operations are left-associative, except the right-associative unary minus operator. The evaluation usually is done
from left to right. Parentheses can change the order of evaluation and some operators have a higher precedence
than others, see section 8.5.17.
The arithmetic operations in HDevelop match the usual definitions. Expressions can have any number of paren-
theses.
The division operator (a1 / a2) can be applied to integer as well as to real. The result is of type real, if at
least one of the operands is of type real. If both operands are of type integer, the division is an integer division.
The remaining arithmetic operators (multiplication, addition, subtraction, and negation) can be applied to either
integer or real numbers. If at least one operand is of type real, the result will be a real number as well.
Examples:
Expression Result
4/3 1
4/3.0 1.3333333
(4/3) * 2.0 2.0
This section describes the operations for bit processing of numbers. The operands have to be integers.
The result of lsh(i1,i2) is a bitwise left shift of i1 that is applied i2 times. If there is no overflow, this is
equivalent to a multiplication by 2i2 . The result of rsh(i1,i2) is a bitwise right shift of i1 that is applied i2
times. For non-negative i1, this is equivalent to a division by 2i2 . For negative i1, the result depends on the used
hardware. For lsh and rsh, the result is undefined if the second operand has a negative value or the value is larger
than 32. More examples can be found in the program bit.hdev.
There are several string operations available to modify, select, and combine strings. Furthermore, some operations
allow you to convert numbers (real and integer) to strings.
$ (string conversion)
See also: tuple_string.
$ converts numbers to strings or modifies strings. The operation has two operands: The first one (left of the $) is
the number that has to be converted. The second one (right of the $) specifies the conversion. It is comparable
to the format string of the printf() function in the C programming language. This format string consists of the
following four parts
<flags><field width>.<precision><conversion>
or as a regular expression:
8.5.8 String Operations 261
Language
strlen(s) length of string tuple_strlen
str_firstn(s,i) cut the first characters of s up to position tuple_str_first_n
i
str_lastn(s,i) cut the characters of s from position i tuple_str_last_n
str_replace(s1,s2,s3) replace substrings of s1 matching s2 tuple_str_replace
with s3
s{i} select character at position i; 0 <= i tuple_str_bit_select
<= strlen(s)-1
s{i1:i2} select substring from position i1 to po- tuple_substr
sition i2
split(s1,s2) split s1 in substrings at s2 tuple_split
join(s1,s2) join substrings in s1 via separator in s2 tuple_join
regexp_match(s1,s2) extract substrings of s1 matching the tuple_regexp_match
regular expression s2
regexp_replace(s1,s2,s3) replace substrings of s1 matching the tuple_regexp_replace
regular expression s2 with s3
regexp_select(s1,s2) select tuple elements from s1 matching tuple_regexp_select
the regular expression s2
regexp_test(s1,s2) return how many tuple elements in s1 tuple_regexp_test
match the regular expression s2
s1 =~ s2 return how many tuple elements in s1 tuple_regexp_test
match the regular expression s2
[-+ #]?([0-9]+)?(\.[0-9]*)?[doxXfeEgGsb]?
which roughly translates to zero or more of the characters in the first bracket pair, followed by zero or more digits,
optionally followed by a dot, which may be followed by digits followed by a conversion character from the last
bracket pair.
Some conversion examples might show it best:
262 HDevelop Language
Input Output
23 $ '10.2f' ' 23.00'
23 $ '-10.2f' '23.00 '
4 $ '.7f' '4.0000000'
1234.56789 $ '+10.3f' ' +1234.568'
255 $ 'x' 'ff'
255 $ 'X' 'FF'
0xff $ '.5d' '00255'
'total' $ '10s' ' total'
'total' $ '-10s' 'total '
'total' $ '10.3' ' tot'
flags Zero or more flags, in any order, which modify the meaning of the conversion specification. Flags may
consist of the following characters:
width An optional string of decimal digits to specify a minimum field width. For an output field, if the converted
value has fewer characters than the field width, it is padded on the left (or right, if the left-adjustment flag -
has been given) to the field width.
precision The precision specifies the minimum number of digits to appear for integer conversions (the field is
padded with leading zeros), the number of digits to appear after the radix character for the e and f conver-
sions, the maximum number of significant digits for the g conversion, or the maximum number of characters
to be printed from a string conversion. The precision takes the form of a period . followed by a decimal
digit string. A null digit string is treated as a zero.
conversion A conversion character indicates the type of conversion to be applied:
d, o, x, X The integer argument is printed in signed decimal (d), unsigned octal (o), or unsigned hexadecimal
notation (x and X). The x conversion uses the numbers and lower-case letters 0123456789abcdef, and
the X conversion uses the numbers and upper-case letters 0123456789ABCDEF. The precision compo-
nent of the argument specifies the minimum number of digits to appear. If the value being converted
can be represented in fewer digits than the specified minimum, it is expanded with leading zeroes. The
default precision is 1. The result of converting a zero value with a precision of 0 is no characters.
f The floating-point number argument is printed in decimal notation in the style [-]ddd.ddd, where the
number of digits after the radix character, ., is equal to the precision specification. If the precision is
omitted from the argument, six digits are output; if the precision is explicitly 0, no radix appears.
e,E The floating-point-number argument is printed in the style [-]d.ddde+dd, where there is one digit
before the radix character, and the number of digits after it is equal to the precision. When the precision
is missing, six digits are produced; if the precision is 0, no radix character appears. The E conversion
character produces a number with E introducing the exponent instead of e. The exponent always
contains at least two digits. However, if the value to be printed requires an exponent greater than two
digits, additional exponent digits are printed as necessary.
g, G The floating-point-number argument is printed in style f or e (or in style E in the case of a G conversion
character), with the precision specifying the number of significant digits. The style used depends on
8.5.8 String Operations 263
the value converted; style e is used only if the exponent resulting from the conversion is less than -4
or greater than or equal to the precision. Trailing zeros are removed from the result. A radix character
appears only if it is followed by a digit.
s The argument is taken to be a string, and characters from the string are printed until the end of the string
or the number of characters indicated by the precision specification of the argument is reached. If the
precision is omitted from the argument, it is interpreted as infinite and all characters up to the end of
the string are printed.
In no case does a nonexistent or insufficient field width cause truncation of a field; if the result of a conversion
is wider than the field width, the field is expanded to contain the conversion result.
Language
Examples for the string conversion can be found in the program string.hdev.
+ (string concatenation)
See also: tuple_add.
The string concatenation (+) can be applied in combination with strings or all numerical types; if necessary, the
operands are first transformed into strings (according to their standard representation). At least one of the operands
has to be already a string so that the operator can act as a string concatenator. In the following example, a file
name for example, 'Name5.tiff', is generated. For this purpose, two string constants ('Name' and '.tiff') and
an integer value (the loop-index i) are concatenated:
for i := 1 to 5 by 1
read_image (Image, 'Name'+i+'.tiff')
endfor
str(r)chr
See also: tuple_strchr, tuple_strrchr.
str(r)chr(s1,s2) returns the index of the first (last) occurrence of one of the character in s2 in string s1, or -1
if none of the characters occur in the string. s1 may be a single string or a tuple of strings.
str(r)str
See also: tuple_strstr, tuple_strrstr.
str(r)str(s1,s2) returns the index of the first (last) occurrence of string s2 in string s1, or -1 if s2 does not
occur in the string. s1 may be a single string or a tuple of strings.
strlen
See also: tuple_strlen.
strlen(s) returns the number of characters in s.
str_firstn
See also: tuple_str_first_n.
str_firstn(s,i) returns the characters from the beginning of string s up to position i.
str_lastn
See also: tuple_str_last_n.
str_lastn(s,i) returns the character from position i of string s to the end.
str_replace
See also: tuple_str_replace.
str_replace(s1,s2,s3) replaces all substrings of s1 that match s2 with s3.
264 HDevelop Language
{}
See also: tuple_str_bit_select, tuple_substr.
s{i} selects a single character (specified by index position) from s. The index ranges from zero to the length of
the string minus 1. The result of the operator is a string of length one.
s{i1:i2} returns all characters from the first specified index position (i1) up to the second specified position (i2)
in s as a string. The index ranges from zero to the length of the string minus 1.
split
See also: tuple_split.
split(s1,s2) divides the string s1 into single substrings. The string is split at those positions where it contains
a character from s2. As an example, the result of
split('/usr/image:/usr/proj/image',':')
['/usr/image','/usr/proj/image']
join
See also: tuple_join.
join(s1,s2) joins the substrings in s1. The separator in s2 is placed between the substrings. As an example, the
result of
['/usr/image:/usr/proj/image', '/usr/image;/usr/proj/image']
Regular Expressions
HDevelop provides string functions that use Perl compatible regular expressions. Detailed information about them
can be found in the Reference Manual at the descriptions of the corresponding operators, which have the same name
but start with tuple_. In particular, at the description of tuple_regexp_match you find further information about
the used syntax, a list of possible options, and a link to suitable literature about regular expressions.
regexp_match
See also: tuple_regexp_match.
regexp_match(s1,s2) searches for elements of the tuple s1 that match the regular expression s2. It returns a tu-
ple with the same size as the input tuple (exceptions exist when working with capturing groups, see the description
of tuple_regexp_match in the Reference Manual for details). The resulting tuple contains the matching results
for each tuple element of the input tuple. For a successful match the matching substring is returned. Otherwise, an
empty string is returned.
regexp_replace
See also: tuple_regexp_replace.
regexp_replace(s1,s2,s3) replaces substrings in s1 that match the regular expression s2 with the string given
in s3. By default, only the first matching substring of each element in s1 is replaced. To replace all occurrences,
the option ’replace_all’ has to be set in s2, see tuple_regexp_replace.
For example:
substitutes file names that look like JPEG images with PNG images.
8.5.9 Set Operations 265
regexp_select
See also: tuple_regexp_select.
regexp_select(s1,s2) returns only the elements of the tuple s1 that match the regular expression s2. In
contrast to regexp_match, the original tuple elements instead of the matching substrings are returned. Tuple
elements that do not match the regular expression are discarded.
For example:
Language
sets Selection to all the strings from List that look like file names of JPEG images. Please note that the
backslash character has to be escaped to be preserved.
regexp_test
See also: tuple_regexp_test.
regexp_test(s1,s2) returns the number of elements of the tuple s1 that match the regular expression s2.
Additionally, a short-hand notation of the operator is available, which is convenient in conditional expressions:
s1 =~ s2
Language
not l negation tuple_not
V1 := [18.8,132.4,33,19.3]
V2 := [233.23,32.786,234.4224,63.33]
Diff := V1 - V2
Distance := sqrt(sum(Diff * Diff))
Dotvalue := sum(V1 * V2)
First, the Euclidean distance of the two vectors V1 and V2 is computed, by using the formula:
sX
d= (V 1i − V 2i )2
i
The difference and the multiplication (square) are successively applied to each element of both vectors. Afterwards
sum computes the sum of the squares. Then the square root of the sum is calculated. After that the dot product of
V1 and V2 is determined by the formula:
X
hV 1, V 2i = (V 1i ∗ V 2i )
i
8.5.16 Miscellaneous Functions 269
Language
deviation(a) standard deviation tuple_deviation
cumul(a) cumulative sums of a tuple tuple_cumul
median(a) median of a tuple tuple_median
select_rank(a,i) element at rank i of a tuple tuple_select_rank
√
sqrt(a) square root a tuple_sqrt
√
cbrt(a) cube root 3 a tuple_cbrt
√
hypot(a,b) hypotenuse a2 + b2 tuple_hypot
deg(a) convert radians to degrees tuple_deg
rad(a) convert degrees to radians tuple_rad
real(a) convert integer to real tuple_real
int(a) truncate real to integer tuple_int
round(a) convert real to integer tuple_round
abs(a) absolute value of a (integer or real) tuple_abs
fabs(a) absolute value of a (always real) tuple_fabs
ceil(a) smallest integer value not smaller than a tuple_ceil
floor(a) largest integer value not greater than a tuple_floor
fmod(a1,a2) fractional part of a1/a2, with the same sign as a1 tuple_fmod
sgn(a) elementwise sign of a tuple tuple_sgn
Table 8.26 shows the provided precedence of the operations for control data.
or
xor
and
bor
bxor
band
== != =~ = #
> < >= <=
[==] [!=] [= ] [=] [#]
[>] [<] [>=] [<=]
+ -
* / %
- (unary minus) not
bnot
$
.
8.6 Vectors
A vector is a container that can hold an arbitrary number of elements, all of which must have the exact same
variable type (tuple, iconic object, or vector). The variable type “vector” is specific to HDevelop. It is available in
HDevelop 12.0 or higher. Please note that programs utilizing vector variables cannot be executed in older versions
of HDevelop.
A vector of tuples or objects is called one-dimensional, a vector of vectors of tuples or objects is two-dimensional,
and so on. The type of a vector must not change within the program. Its dimension has to remain constant, and
vectors of tuples must not be assigned iconic objects or vice versa.
This is the definition of a vector in EBNF (Extended Backus-Naur Form) grammar:
Construction of Vectors
A vector is defined by providing a comma-separated list of its elements in curly brackets.
This is equivalent to
t1 := 1
vectorT := {t1, t1 * 2, 3}
Language
connection (Region, ConnectedRegions)
vectorO := {Image, Region, ConnectedRegions}
It is also possible to define vector variables using the .at() and .insert() operations.
The list of the vector’s elements may also be empty, that is, an empty vector {} is valid as in the following example:
Note, however, that an empty vector has no specific type. Therefore, all of the following three empty assignments
are valid:
vectorO2 := vectorO
vectorT2 := vectorT
vectorV2 := vectorV
vectorO2 := {}
vectorT2 := {}
vectorV2 := {}
Assigning an empty vector to a vector variable is equivalent to the .clear() operation. This means that the
assignment of an empty vector to a variable is not sufficient to define the variable’s type, see section 8.3.1 on page
251. Such a variable will have an undefined type (and, therefore, be invalid) unless its type is defined properly
elsewhere in the program.
The .at() operation is also used to set vector elements. Writing to a non-existing vector element is allowed. If
necessary, the vector is automatically filled with empty elements.
The at() operation also allows you to construct a vector dynamically in a loop:
272 HDevelop Language
for i:= 0 to 5 by 1
vecT.at(i) := gen_tuple_const(i,5)
endfor
The .insert() operation specifies an index position and a value. It shifts the values from the given index to the
end by one position, and sets the value at the index to the new value.
The .remove() operation performs the opposite operation. It removes the value at the specified position and
moves all following values to the left.
Like with the .at() operation, the vector is automatically filled with empty elements if necessary.
The .concat() operation concatenates two vectors of the same type and dimension.
i := vectorT.length() // i := 3
j := vectorV.length() // j := 2
k := vectorV.at(0).length() // k := 3
vectorT.clear()
* vectorT := vectorO // illegal: vectorT is a tuple vector
* vectorT := vectorV // illegal: vectorT is one-dimensional
v := {1, 2, 3}
v.insert(1, 5).insert(2, 4).remove(0) // sets v to {5, 4, 2, 3}
Use the special operator executable_expression to enter a modifying vector operation into the operator win-
dow.
The convenience operator convert_tuple_to_vector_1d stores the elements of the input tuple as single ele-
ments of the one-dimensional output vector.
Language
convert_tuple_to_vector_1d (T, 1, V) // V := {[1],[2],[3],[4],[5],[6],[7],[8],[9]}
8.7 Dictionaries
A dictionary (often also called “dict”) is an associative, array-like container. Similarly to vectors, it can store an
arbitrary number of values. In contrast, the values are each associated with unique keys (integers or strings), and
each key can refer either to a tuple or to an iconic object.
For example:
create_dict (DictHandle)
* Add data.
set_dict_tuple (DictHandle, 'SingleInteger', 27)
set_dict_tuple (DictHandle, 'MixedTuple', ['The answer', 41])
set_dict_object (Image, DictHandle, 'SingleImage')
* Access by key.
get_dict_tuple (DictHandle, 'MixedTuple', Tuple)
* Update value.
RightAnswer := Tuple[1] + 1
set_dict_tuple (DictHandle, 'MixedTuple', ['The answer',RightAnswer])
Read dict To look up the value associated with a key, use the dict access expression.
<DictValue> := <DictHandle>.['<key>']
Write dict To assign a value to a key, use a dict access expression. If the key is not present in the dictionary yet,
the corresponding key-value pair will be added.
<DictHandle>.['<key>'] := <NewDictValue>
The following code shows the example from above using dict expressions:
274 HDevelop Language
* Add data.
DictHandle := dict{['SingleInteger']: 27, ['MixedTuple']: ['The answer', \
41]}
read_image (DictHandle.['SingleImage'], 'printer_chip/printer_chip_01')
* Access dict's tuple element.
RightAnswer := DictHandle.['MixedTuple'][1] + 1
* Update dict value.
DictHandle.['MixedTuple'] := ['The answer',RightAnswer]
Note that the dict access expression can increase the runtime when handling larger tuples, because the tuple is
copied internally. Therefore, instead of repeatedly accessing a tuple in a loop, it is more efficient to access the
tuple outside the loop and subsequently iterate over the tuple content.
Simplifying the code using static syntax
Often, you can further simplify the expressions compared to “dynamic syntax” and use “static syntax” instead. To
create a dictionary in static syntax notation, use
<DictHandle> := dict{<key>: <value>}
For example:
DictHandle := dict{SingleInteger: 27, MixedTuple: ['The answer', 42]}
The example from above now looks like this:
* Add data.
DictHandle := dict{SingleInteger: 27, MixedTuple: ['The answer', 41]}
read_image (DictHandle.SingleImage, 'printer_chip/printer_chip_01')
* Access dict's tuple element.
RightAnswer := DictHandle.MixedTuple[1] + 1
* Update dict value.
DictHandle.MixedTuple := ['The answer',RightAnswer]
ImagePath := 'printer_chip/printer_chip_01'
read_image (Image, ImagePath)
Params := dict{[ImagePath]: 1}
Here, ImagePath is a variable and used as the key in the dictionary assignment:
<key> = 'printer_chip/printer_chip_01').
We can then extend the example like this:
ImagePathPrefix := 'printer_chip/printer_chip_0'
for Id := 1 to 5 by 1
read_image (Image, ImagePathPrefix + Id)
Params := dict{[ImagePathPrefix + Id]: Id}
endfor
ImagePathPrefix := 'printer_chip/printer_chip_0'
for Id := 1 to 5 by 1
read_image (Image, ImagePathPrefix + Id)
Params := dict{ImagePathPrefix + Id: Id} // error!
endfor
In this case, ImagePathPrefix is interpreted as a string and the dictionary Params would contain the syntactically
wrong key 'ImagePathPrefix' + 1, while the output of read_image is not being used any further.
Language
The following example shows another case requiring dynamic syntax:
for i := 0 to |Keys| - 1 by 1
Val := Params.[Keys[i]]
endfor
The same could be written with a single expression. When using multiple keys in one dict access or assignment
expression, the elements to which the keys refer must all be tuples of length 1.
Keys := []
* Vals will always be [] in this case.
Vals := Params.[Keys]
Nesting dictionaries
The dict init expression allows arbitrary nesting. You can create further dictionaries within a dictionary. For
example:
Restrictions
Currently, the following restrictions apply in regard to dict expressions:
Objects The copy is a reference, as in copy_obj. In particular, changes made using the operators set_grayval
or overpaint_region affect the object stored in the dictionary as well.
Handles Storing any handle in the dictionary will copy the handle value, but not the resource behind the handle.
This is illustrated in the following example:
create_generic_shape_model (ModelID1)
set_generic_shape_model_param (ModelID1, 'min_score', 0.42)
DictHandle := dict{sbm_model_handle: ModelID1}
ModelID2 := DictHandle.sbm_model_handle
set_generic_shape_model_param (ModelID1, 'min_score', 0.73)
get_generic_shape_model_param (ModelID2, 'min_score', MinScoreModel2)
ModelID1 and ModelID2 are both handles for the same model. As a consequence, the retrieved
MinScoreModel2 is 0.73, the value set a line above and not the value set before setting the model han-
dle as dictionary entry.
The identifiers listed in table 8.27 on page 278 are reserved words and their usage is strictly limited. Here are some
examples where they cannot be used:
• As variable names.
global := 55 // doesn't work!
• As key in static dictionary expressions.
X := dict{global: 55} // doesn't work!
• As thread id in par_start.
par_start<global>: tuple_div(3, 2, Out) // doesn't work!
The operators introduced in this section execute a block of operators conditionally or repeatedly. Usually, these
operators come in pairs: One operator marks the start of the block while the other marks the end. The code lines
in-between are referred to as the body of a control flow structure.
When you enter a control flow operator to start a block, HDevelop also adds the corresponding closing operator
by default to keep the program code balanced. In addition, the IC is placed between the control flow operators.
This is fine for entering new code blocks. If you want to add control flow operators to existing code, you can also
add the operators individually. Keep in mind, however, that a single control flow operator is treated as invalid code
until its counterpart is entered as well.
In the following, <condition> is an expression that evaluates to an integer or boolean value. A condition is
false if the expression evaluates to 0 (zero). Otherwise, it is true. HDevelop provides the following operators to
control the program flow:
8.9 Control Flow Operators 277
if ... endif This control flow structure executes a block of code conditionally. The operator if takes a condition
as its input parameter. If the condition is true, the body is executed. Otherwise the execution is continued at
the operator call that follows the operator endif.
To enter both if and endif at once, select the operator if in the operator window and make sure the check
box next to the operator is ticked.
if (<condition>)
...
endif
Language
if ... else ... endif Another simple control flow structure is the condition with alternative. If the condition is
true, the block between if and else is executed. If the condition is false, the part between else and endif
is executed.
To enter all three operators at once, select the operator ifelse in the operator window and make sure the
check box next to the operator is ticked.
if (<condition>)
...
else
...
endif
278 HDevelop Language
elseif This operator is similar to the else-part of the previous control flow structure. However, it allows you
to test for an additional condition. The block between elseif and endif is executed if <condition1> is
false and <condition2> is true. elseif may be followed by an arbitrary number of additional elseif
instructions. The last elseif may be followed by a single else instruction.
if (<condition1>)
...
elseif (<condition2>)
...
endif
Language
This is syntactically equivalent and thus a shortcut for the following code block:
if (<condition1>)
...
else
if (<condition2>)
...
endif
endif
while ... endwhile This is a looping control flow structure. As long as the condition is true, the body of the
loop is executed. In order to enter the loop, the condition has to be true in the first place. The loop can be
restarted and terminated immediately with the operator continue and break..
To enter both while and endwhile at once, select the operator while in the operator window and make
sure the check box next to the operator is ticked.
while (<condition>)
...
endwhile
repeat ... until This loop is similar to the while loop with the exception that the condition is tested at the end
of the loop. Thus, the body of a repeat ... until loop is executed at least once. Also in contrast to the
while loop, the loop is repeated if the condition is false, until it is finally true.
To enter both repeat and until at once, select the operator until in the operator window and make sure
the check box next to the operator is ticked.
repeat
...
until (<condition>)
for ... endfor The for loop is controlled by a start and an end value and an increment value, step, that determines
the number of loop steps. These values may also be expressions, which are evaluated immediately before
the loop is entered. The expressions may be of type integer or of type real. If all input values are of type
integer, the loop variable will also be of type integer. In all other cases the loop variable will be of type
real.
Please note that the for loop is displayed differently in the program window than entered in the operator
window. What you enter in the operator window as for(start,end,step,index ) is displayed in the
program window as:
To enter both for and endfor at once, select the operator for in the operator window and make sure the
check box next to the operator is ticked.
The start value is assigned to the index variable. The loop is executed as long as the following conditions are
true: 1) The step value is positive, and the loop index is smaller than or equal to the end value. 2) The step
280 HDevelop Language
value is negative, and the loop index is greater than or equal to the end value. After a loop cycle, the loop
index is incremented by the step value and the conditions are evaluated again.
Thus, after executing the following lines,
for i := 1 to 5 by 1
j := i
endfor
for i := 5 to 1 by -1
j := i
endfor
old_x := 0
old_y := 0
dev_set_color ('red')
dev_set_part(0, 0, 511, 511)
for x := 1 to 511 by 1
y := sin(x / 511.0 * 2 * 3.1416 * 3) * 255
disp_line (WindowID, -old_y+256, old_x, -y+256, x)
old_x := x
old_y := y
endfor
In this example the assumption is made that the window is of size 512×512. The drawing is always done
from the most recently evaluated point to the current point.
continue The operator continue forces the next loop cycle of a for, while, or repeat loop. The loop condition
is tested, and the loop is executed depending on the result of the test.
In the following example, a selection of RGB color images is processed. Images with channel numbers other
than three are skipped through the use of the operator continue. An alternative is to invert the condition
and put the processing instructions between if and endif. But the form with continue tends to be much
more readable when very complex processing with lots of lines of code is involved.
i := |Images|
while (i)
Image := Images[i]
count_channels (Image, Channels)
if (Channels != 3)
continue
endif
* extensive processing of color image follows
endwhile
break The operator break enables you to exit for, while, and repeat loops. The program is then continued at
the next line after the end of the loop.
A typical use of the operator break is to terminate a for loop as soon as a certain condition becomes true,
for example, as in the following example:
8.9 Control Flow Operators 281
Number := |Regions|
AllRegionsValid := 1
* check whether all regions have an area <= 30
for i := 1 to Number by 1
ObjectSelected := Regions[i]
area_center (ObjectSelected, Area, Row, Column)
if (Area > 30)
AllRegionsValid := 0
break ()
endif
Language
endfor
In the following example, the operator break is used to terminate an (infinite) while loop as soon as one
clicks into the graphics window:
while (1)
grab_image (Image, FGHandle)
dev_error_var (Error, 1)
dev_set_check ('~give_error')
get_mposition (WindowHandle, R, C, Button)
dev_error_var (Error, 0)
dev_set_check ('give_error')
if ((Error = H_MSG_TRUE) and (Button != 0))
break ()
endif
endwhile
switch ... case ... endswitch The switch block allows you to control the program flow via a multiway branch.
The branch targets are specified with case statements followed by an integer constant. Depending on an
integer control value the program execution jumps to the matching case statements and continues to the next
break statement or the closing endswitch statement. An optional default statement can be defined as the
last jump label within a switch block. The program execution jumps to the default label if no preceding
case statement matches the control expression.
...
switch (Grade)
case 1:
Result := 'excellent'
break
case 2:
Result := 'good'
break
case 3:
Result := 'acceptable'
break
case 4:
case 5:
Result := 'unacceptable'
break
default:
Result := 'undefined'
endswitch
...
stop The operator stop stops the program after the operator is executed. The program can be continued by
pressing the Step Over or Run button.
exit The exit operator terminates the HDevelop session.
return The operator return returns from the current procedure call to the calling procedure. If return is called
in the main procedure, the PC jumps to the end of the program, and the program is finished.
282 HDevelop Language
try ... catch ... endtry This control flow structure enables dynamic exception handling in HDevelop. The pro-
gram block, runtime errors, between the operators try and catch is watched for exceptions. If an exception
occurs, diagnostic data about what caused the exception is stored in an exception tuple. The exception tuple
is passed to the catch operator, and program execution continues from there. The program block between
the operators catch and endtry is intended to analyze the exception data and react to it accordingly. If no
exception occurs, this program block is never executed.
See section 8.10, and the reference manual, for example, the operator try for detailed information.
throw The operator throw allows you to generate user-defined exceptions.
dev_set_check('~give_error')
HDevelop will then ignore any errors in the program. Consequently, the programmer has to take care of the
error handling. Every operator call provides a return value (or error code) which signals success or failure of its
execution. This error code can be accessed through a designated error variable:
dev_error_var(ErrorCode, 1)
This operator call instantiates the variable ErrorCode. It stores the error code of the last executed operator. Using
this error code, the program can depend its further flow on the success of an operation.
...
if (ErrorCode != H_MSG_TRUE)
* react to error
endif
* continue with program
...
The error message related to a given error code can be obtained with the operator get_error_text. This is useful
when reporting errors back to the user of the program.
If the error is to be handled in a calling procedure, an appropriate output control variable has to be added to the
interface of each participating procedure, or the error variable has to be defined as a global variable (see section
8.3.2).
HDevelop supports dynamic exception handling, which is comparable to the exception handling in C++ and C#.
A block of program lines is watched for run-time errors. If an error occurs, an exception is raised and an associated
exception handler is called. An exception handler is just another block of program lines, which is invisible to
the program flow unless an error occurs. The exception handler may directly act on the error or it may pass the
associated information (the exception) on to a parent exception handler. This is also known as rethrowing an
exception.
In contrast to the tracking method described in the previous section, the exception handling requires HDevelop to
Language
be set up to stop on errors. This is the default behavior. It can also be turned on explicitly:
dev_set_check('give_error')
Furthermore, HDevelop can be configured to let the user choose whether or not an exception is thrown, or to throw
exceptions automatically. This behavior is set in the preferences tab General Options . Experienced User.
An HDevelop exception is a tuple containing data related to a specific error. It always contains the error code as
the first item. The operator dev_get_exception_data provides access to the elements of an exception tuple.
HDevelop exception handling is applied in the following way:
...
try
* start block of watched program lines
...
catch(Exception)
* get error code
ErrorCode := Exception[0]
* react to error
endtry
* program continues normally
...
To start a new thread, prefix the corresponding operator or procedure call with the par_start qualifier:
This call starts the hypothetical procedure gather_data() as a new subthread in the background and continues to
execute the subsequent program lines. The thread ID is returned in the variable ThreadID, which must be specified
in angle brackets. Unlike in HDevelop, in HDevEngine a given ThreadID is only valid within the procedure that
started the thread.
Note that par_start is not an actual operator but merely a qualifier that modifies the calling behavior. Therefore,
it is not possible to select par_start in the operator window.
If starting a new subthread would exceed the configured maximum number of threads, see above, an exception is
thrown.
You can also start procedure or operator calls as a subthread from the operator window, see figure 8.3. To do this,
open the section Advanced Parallelization Options at the bottom of the operator window, tick the check
box, and enter the name of the variable that will hold the thread ID. If you double-click a program containing
the par_start qualifier, the parallelization options will also be displayed in the operator window. For certain
program lines, comments, declarations, loops, or assignments, par_start is not supported and the corresponding
options will also not be available in the operator window. For a general description of the operator window, see
section 6.11 on page 91.
It is supported to start multiple threads in a loop. In that case, the thread IDs need to be collected so that all threads
can be referenced later:
ThreadIDs := []
for Index := 1 to 5 by 1
par_start <ThreadID> : gather_data()
ThreadIDs := [ThreadIDs, ThreadID]
endfor
It is often more convenient to collect the thread IDs in a vector variable (page 270):
for Index := 1 to 5 by 1
par_start <ThreadIDs.at(Index - 1)> : gather_data()
endfor
for Index := 1 to 5 by 1
par_start <ThreadIDs.at(Index - 1)> : gather_data(Result) // BEWARE!!!
endfor
In the example, all the threads would return their result in the same variable which is certainly not what was
intended. The final value of Result would be the unpredictable return value of the thread that finishes last, and all
other results would be lost.
An easy solution to this problem is to collect the returned data in a vector variable as shown previously with the
thread IDs:
Language
for Index := 1 to 5 by 1
par_start <ThreadIDs.at(Index - 1)> : gather_data(Result.at(Index - 1))
endfor
Here, each invocation of gather_data returns its result in a unique slot of the vector variable Result.
Use the operator par_join to wait for the completion of a single thread or a group of threads.
As an example why this is necessary suppose we want to call a procedure that performs some magic calculation
in the background and returns a count number as a result. In the subsequent program lines we want to use that
number for further calculations.
Relying on the subthread to be fast enough is most likely going to fail. Therefore, an explicit call to par_join is
required beforehand.
Note that in HDevelop it is not strictly required to use par_join because the main thread will always outlive the
subthreads. However, omitting it might lead to trouble if the program is going to be executed in HDevEngine,
see Programmer’s Guide, page 140 or exported to a programming language. Similarly, access to global variables
might need some additional synchronization if the program is going to be exported.
Given the example from the previous section, waiting for the finishing of all the threads that were started in a loop
is achieved using the following lines.
convert_vector_to_tuple(ThreadIDs, Threads)
par_join(Threads)
Please note that the thread IDs had been collected in a vector variable. Thus, the conversion to a tuple is necessary
for par_join to work properly.
The par_join operator blocks the further execution of the procedure it has been called from until all specified
threads have finished. In the subsequent program lines, results from the corresponding threads can then be accessed
reliably.
286 HDevelop Language
In
general, threads in HDevelop are only executed in parallel when the program runs continuously after pressing
F5 . In all other execution modes only the selected thread is started and all other threads remain stopped unless
an explicit user interaction advances their execution. Active breakpoints, stop instructions, runtime errors or
uncaught exceptions also cause all threads to stop so their current state can be evaluated. This convention enables a
clearly defined debugging process because it eliminates uncontrollable side-effects from other threads. Any editing
action in the program window will also cause a concurrently running program to stop.
Threads cannot be “killed” externally. They can be stopped between operator calls or by aborting interruptible
operators. If any thread executes a long-running operator that cannot be interrupted at the time HDevelop tries to
stop the program execution, a corresponding message will be displayed in the status line, and the corresponding
thread will eventually stop after the operator has finished.
Selected Thread
Exactly one of the threads is the so-called selected thread; by default this corresponds to the main thread of the
program. The position of the PC, the status of the call stack, and the state of the variables in the variable window
are linked to the selected thread. The selected thread can change automatically to a thread that stops, for example,
by a breakpoint, stop instruction, an uncaught exception, or a draw operator.
How to select a specific thread see section 8.11.4. All run modes other then continuous execution apply only to the
selected thread. Program lines unrelated to the selected thread will be grayed out in the program window.
Thread Lifetime
A thread exists as long as it is still being referenced by a variable even if its execution has finished. This is necessary
to reference the thread in a par_join instruction, or to set back the PC of the corresponding thread for debugging.
However, the lifetime of a thread ends if the PC is manually set back to a program line before its invocation. Apart
from that the lifetime of all subthreads ends when the program is reset using F2 . During its lifetime a thread is
listed in the Thread View / Call Stack window from where it can be selected and managed.
Finished, but still “living” threads might cause the configured thread number limit to be exceeded unintentionally
when new threads are started at a later time. Also, it can have a negative effect on the run time behavior if finished
threads are “killed” at the same time new threads are being executed. To explicitly “kill” finished threads, it is
sufficient to reset the variables that are referencing their thread IDs as in the following example.
for Index := 0 to 4 by 1
par_start <ThreadIDs.at(Index)> : gather_data()
endfor
...
convert_vector_to_tuple (ThreadIDs, Threads)
par_join (Threads)
...
ThreadIDs := {}
Threads := []
Error Handling
Each thread can specify its own error handling, for example, using dev_set_check('~give_error'). New
subthreads inherit the error handling mode from their parent thread. Exception handling using try .. catch works
only within a thread. In the main thread it is not possible to catch an exception that is thrown in a subthread.
8.11.4 Inspecting Threads 287
The current execution status of the program and its threads is displayed in the combined thread view/call stack
window. Select Execute . Thread View / Call Stack or click in the tool bar (see also Thread View /
Call Stack (page 160)). The upper half of the window lists all existing threads while the lower half displays
the call stack of the selected thread. To illustrate the interaction with this window, consider the following (silly)
example.
for Index:= 1 to 5 by 1
Language
par_start <ThreadIDs.at(Index - 1)> : wait_seconds(Index)
endfor
wait_seconds(2)
stop()
After pressing F5 , the program will start five subthreads, and eventually reach the stop instruction, leaving some
subthreads still running while others will already have finished. The corresponding thread view is displayed in
figure 8.4. Note that the unfinished threads are in a stopped state because of another thread (in this case caused by
the stop instruction in the main thread).
The thread view lists the properties of all threads in a table. The status icon in the first column of each thread
visualizes the current execution state. The currently selected thread is marked by the arrow in the status icon
and is also highlighted in bold text. The other five threads are the subthreads started from the main thread. To select
another thread, double-click it in the thread view. This will also update the PC, the call stack, and the variables
according to the selected thread. The active procedure of the selected thread will be displayed in the program
window.
The meaning of the columns of the thread view is as follows:
Column Description
Thread ID The unique number assigned to the thread when it is started.
State The current execution state of the thread.
Run Mode The mode that the thread was started in the last time.
Stop Mode The reason for stopping the thread.
Caller The position (procedure and line number) from where the thread was
started.
References (hidden by default) The number of references to this thread.
288 HDevelop Language
Notification line
Thread view window
Threads can explicitly be suspended and resumed in the thread view window. Suspending a thread only works
between operator calls. If the thread is currently running, the current operator will still be executed before the
thread is frozen.
To suspend a thread, right-click the thread entry and select Suspend Thread. Suspended threats are “frozen” in
their current state and will defer subsequent run commands until the threads are resumed again. Run commands
will change the run status of the suspended threads but the actual execution will be prevented.
To resume a suspended thread again, right-click the thread entry and select Resume Thread.
Language
tuple creation [i1:i2:i3] tuple_gen_sequence
tuple creation [i1:i2] tuple_gen_sequence
tuple creation gen_tuple_const(i1,i2) tuple_gen_const
repeat tuple rep(s,i) tuple_repeat
repeat tuple elements rep_elem(s,i) tuple_repeat_elem
division a1 / a2 tuple_div
multiplication a1 * a2 tuple_mult
modulo a1 % a2 tuple_mod
addition a1 + a2 tuple_add
subtraction a1 - a2 tuple_sub
negation -a tuple_neg
left shift lsh(i1,i2) tuple_lsh
right shift rsh(i1,i2) tuple_rsh
bitwise and i1 band i2 tuple_band
bitwise xor i1 bxor i2 tuple_bxor
bitwise or i1 bor i2 tuple_bor
bitwise complement bnot i tuple_bnot
string conversion v$s tuple_string
string concatenation v1 + v2 tuple_add
search character strchr(s1,s2) tuple_strchr
search character (reverse) strrchr(s1,s2) tuple_strrchr
search string strstr(s1,s2) tuple_strstr
search string (reverse) strrstr(s1,s2) tuple_strrstr
length of string strlen(s) tuple_strlen
first characters str_firstn(s,i) tuple_str_first_n
last characters str_lastn(s,i) tuple_str_last_n
select character s{i} tuple_str_bit_select
select substring s{i1:i2} tuple_str_bit_select
split string split(s1,s2) tuple_split
join strings join(s1,s2) tuple_join
replace strings str_replace(s1,s2,s3) tuple_str_replace
regular expression match regexp_match(s1,s2) tuple_regexp_match
regular expression replace regexp_replace(s1,s2,s3) tuple_regexp_replace
regular expression select regexp_select(s1,s2) tuple_regexp_select
290 HDevelop Language
Language
power function pow(a1,a2) tuple_pow
ldexp function ldexp(a1,a2) tuple_ldexp
gamma function tgamma(a) tuple_tgamma
logarithm of the absolute value of the gamma function lgamma(a) tuple_lgamma
error function erf(a) tuple_erf
complementary error function erfc(a) tuple_erfc
minimum min(t) tuple_min
elementwise minimum min2(t1,t2) tuple_min2
maximum max(t) tuple_max
elementwise maximum max2(t1,t2) tuple_max2
sum function sum(t) tuple_sum
mean value mean(a) tuple_mean
standard deviation deviation(a) tuple_deviation
cumulative sum cumul(a) tuple_cumul
median median(a) tuple_median
element rank select_rank(a,i) tuple_select_rank
square root sqrt(a) tuple_sqrt
cube root cbrt(a) tuple_cbrt
hypotenuse hypot(a,b) tuple_hypot
radians to degrees deg(a) tuple_deg
degrees to radians rad(a) tuple_rad
integer to real real(a) tuple_real
real to integer int(a) tuple_int
real to integer round(a) tuple_round
absolute value abs(a) tuple_abs
floating absolute value fabs(a) tuple_fabs
ceiling function ceil(a) tuple_ceil
floor function floor(a) tuple_floor
fractional part fmod(a1,a2) tuple_fmod
elementwise sign sgn(a) tuple_sgn
sort elements sort(t) tuple_sort
sort elements (returns index) sort_index(t) tuple_sort_index
reverse element order inverse(t) tuple_inverse
292 HDevelop Language
Language
21034 Password protection/locked procedures
21035 Procedures with advanced language elements are not supported for the selected action.
21035 Parallel execution statements, iconic assignments, or iconic comparisons
21036 Procedures with vector variables are not supported for the selected action.
21036 Vector variables
21040 Unable to open file
21041 Unable to read from file
21042 Unable to write to file
21043 Unable to rename file
21044 Unable to open file: invalid file name
21050 For this operator the parallel execution with par_start is not supported
21051 Thread creation failed
21052 Thread creation failed: exceeded maximum number of subthreads
21060 Iconic variable is not instantiated
21061 Control variable is not instantiated (no value)
21062 Wrong number of control values
21063 Wrong value type of control parameter
21064 Wrong value of control parameter
21065 Control parameter does not contain a variable
21066 Control parameter must be a constant value
21067 Wrong number of control values in condition variable
21068 Wrong type: Condition variable must be an integer or boolean
21070 Variable names must not be empty
21071 Variable names must not start with a number
21072 Invalid variable name
21073 Invalid name for a control variable: the name is already used for an iconic variable
21074 Invalid name for an iconic variable: the name is already used for a control variable
21075 An iconic variable is used in the wrong context: a control variable or a vector is expected
21076 A control variable is used in the wrong context: an iconic variable or a vector is expected
21077 An iconic vector variable is used in the wrong context: a control variable or a single value
is expected
21078 A control vector variable is used in the wrong context: an iconic variable or a single value
is expected
21080 For loop variable must be a number
294 HDevelop Language
Language
21133 Right square bracket ’]’ expected
21134 Unmatched right parenthesis ’)’ found
21135 Unmatched right curly brace ’}’ found
21136 Unmatched right square bracket ’]’ found
21137 Second bar ’|’ expected
21138 Function name expected before parentheses
21139 Unterminated string detected
21140 Invalid character in an expression identifier detected
21141 Parameter expression expected
21142 Parameter expression is not executable
21143 method or key expected after .
21144 Modifying vector methods are not allowed within parameters
21200 Syntax error in operator expression
21201 Identifier (operator or variable name) expected
21203 Syntax error in parameter list
21204 Parenthesis expected
21205 No parenthesis expected
21206 List of parameters in parenthesis expected
21207 Wrong number of parameters
21208 Unexpected characters at end of line
21209 Assign operator ’:=’ expected
21210 Expression after assign operator ’:=’ expected
21211 Expression in brackets ’[ ]’ for the assign_at index expected
21212 In for statement, after keyword ’by’ expression for parameter ’Step’ expected
21213 In for statement, after keyword ’to’ expression for parameter ’End’ expected
21214 In for statement, after assign operation (’:=’) expression for parameter ’Start’ expected
21215 In for statement, after ’for .. := .. to ..’ keyword ’by’ expected
21216 In for statement, after ’for .. := ..’ keyword ’to’ expected
21217 In for statement, assign operation ’:=’ for initializing the index variable expected
21218 After ’for’ keyword, assignment of ’Index’ parameter expected
21219 In for statement, error after ’by’ keyword in expression of parameter ’Step’
21220 In for statement, error after ’to’ keyword in expression of parameter ’End’ or the following
’by’ keyword
296 HDevelop Language
21221 In for statement, error after assignment operation (’:=’) in expression of parameter ’Start’
or the following ’to’ keyword
21222 In for statement, invalid variable name in parameter ’Index’ or error in the following assign-
ment operation (’:=’)
21223 for statement not complete
21224 In for statement, space after ’for’ expected
21225 In for statement, space after ’to’ expected
21226 In for statement, space after ’by’ expected
21227 This expression cannot be used as loop index
21228 Wrong type: The switch statement requires an integer value as parameter
21229 Wrong type: The case statement requires a constant integer value as parameter
21230 At the end of the case and the default statement a colon is expected
21231 Unknown operator or procedure
21232 Qualifier ’par_start’ before ’:’ or ’<ThreadID>’ expected
21233 ’<ThreadID>’ variable in angle brackets after ’par_start’ expected
21234 ThreadID variable after ’par_start<’ expected
21235 Closing angle bracket (’>’) after ’par_start<ThreadID’ expected
21236 Colon (’:’) after ’par_start<ThreadID>’ expected
21237 Operator or procedure call after ’par_start :’ expected
21238 This expression cannot be used as ThreadID
21239 Dictionary expression cannot be used as parameter in par_start call
21900 Internal value has the wrong type
21901 Internal value is not a vector
21902 Index into internal value is out of range
21903 Internal value is not instantiated
22000 Internal operation in expression failed
22001 Internal operation in a constant expression failed
22010 Parameters are tuples with different size
22011 Division by zero
22012 String exceeds maximum length
22100 Parameter is an empty tuple
22101 Parameter has more than one single value
22102 Parameter is not a single value
22103 Parameter has the wrong number of elements
22104 Parameter contains undefined value(s)
22105 Parameter contains wrong value(s)
22106 Parameter contains value(s) with the wrong type
22200 First parameter is an empty tuple
22201 First parameter has more than one single value
22202 First parameter is not a single value
8.14 Emergency Backup 297
Language
22302 Second parameter is not a single value
22303 Second parameter has the wrong number of elements
22304 Second parameter contains undefined value(s)
22305 Second parameter contains wrong value(s)
22306 Second parameter contains value(s) with the wrong type
22400 Calling context was not set
22401 Accessing an invalid calling context
22402 Error while accessing calling context data
22500 Communication with external application failed
22501 Debug session with external application no longer valid
22502 An unexpected error occurred in the external application
22503 Wrong password to unlock procedure in external application
23100 The generic parameter value is unknown
23101 The generic parameter name is unknown
30000 User defined exception
In case HDevelop ever crashes during program execution, the current program is saved in a temporary location.
After restarting HDevelop, you can restore the backup file to continue your application. You can find the backup
file highlighted in red under your Recent Programs list in the start dialog.
When restarting the computer, the locally saved backup files will be deleted.
The exact location of the data depends on the operating system you are using:
!
Chapter 9
Remote Debugging
Remote Debugging
HDevelop supports the debugging of HDevelop code in stand-alone applications. This is referred to as “remote
debugging”. An external application can execute HDevelop procedures using HDevEngine. If it explicitly enables
remote debug connections, HDevelop may attach to this application. Once the connection has been established,
HDevelop operates in a special debugging mode, see figure 9.1. How to enable remote debugging is described in
the Programmer’s Guide, section 25.2 on page 203. The following sections describe the HDevelop side of remote
debugging.
9.1 Requirements
To be able to use remote debugging, both the external application and HDevelop must be based on the same
version of HALCON. Furthermore, the version of HDevelop must at least be as “large” as that of HDevEngine. If
the external application uses HDevEngine XL, you cannot use HDevelop non-XL.
Please note that cross-platform debugging is supported. You can debug an external application running on Windows
using HDevelop on Linux.
External Application
attach to process
HDevEngine
Procedure A
Procedure B
...
Procedure Z
• Enter the host name and connection port, and, optionally, the password. See figure 9.2. The default host
name (localhost) can be used if both the external application and HDevelop run on the same machine.
Please note that your firewall must be configured to allow connections on the specified port. The default port
number is 57786, but this can be changed in the external application. If a debug session is already active on the
given machine, you can choose to take over the existing session. This will terminate the previous session.
After connecting, large parts of the user interface of HDevelop are grayed out. This applies to the functionality
related to the normal operation of HDevelop, such as procedures, operators, assistants, and preferences. The special
debug mode is indicated in the window title and by the red background color of the status bar, see figure 9.3. The
background color of the program window also changes to indicate the read-only nature of the remote debug mode.
This color can be customized in the preferences in normal mode.
9.3 Debugging 301
9.3 Debugging
Remote debugging is in most parts similar to “local” debugging. The procedure execution is controlled by execu-
tion commands (for example, Stop, Step Over, Step Into etc.). Debugging is only possible in stopped mode.
If HDevEngine is in run mode, the status bar reads “Running... External application”, the following applies:
• No PC is displayed.
• Variables are uninitialized.
• Application threads cannot be selected, and the call stack is empty.
Independent of whether run or stop mode is active, the loaded procedures can be viewed in the program window.
As usual, breakpoints may be set on program lines and variables. Activated breakpoints persist as long as the
external application is running. In contrast, deactivated breakpoints disappear on disconnect. HDevEngine stops
at breakpoints only if HDevelop is actually attached.
Remote Debugging
Pressing Stop causes HDevelop to wait for the external application to execute code in HDevEngine. In stop mode
you can execute program lines step-by-step and view intermediate results in both the variable and the graphics
window. Other than that, the execution cannot be influenced:
If the application executes procedure calls in the GUI thread, attaching a debugger and stopping the execution will
freeze the application GUI.
The program window provides access to all procedures loaded by HDevEngine during the debugging session.
Unloading procedures will not cause the corresponding procedures to be removed from the list. This is to ensure
usability even if the application loads/unloads procedures in a loop. You can disconnect and then reconnect the
debugger to refresh the list.
The external application can load multiple copies of the same procedure. In HDevelop however, duplicate pro-
cedure names are shown only once. As a result, breakpoints set in a duplicate procedure will apply to all loaded
copies. If another copy is loaded at a later time, the breakpoint will not be active for it.
Debugging protected procedures is also supported in HDevelop. If such a procedure is selected in the program
window, you will be prompted for a password to unlock it. Unlocking protected procedures only has an effect
inside HDevelop. The state of the procedures in the external application does not change. The unlocked state is
only valid for the current debug session. If you or another user reconnects to the same process, the procedures
will be locked again. There is no explicit way to relock an unlocked procedure during debugging. You can stop
debugging and attach again to achieve this.
The behavior of handling errors in the HDevelop code of the external application depends on whether or not the
exception is handled in the program code. If the exception is unhandled, the error message will be shown in
HDevelop, and the application will wait for a Step or Run signal before raising HDevEngineException itself. If
the exception is handled, there is no way to stop the execution.
302 Remote Debugging
9.7 Threading
Application threads are supported. After sending a stop signal to the application (Stop or F9 ), a random thread
will be selected in HDevelop. Stepping over the last line of a procedure switches threads if another stopped HDev-
Engine thread is waiting. In case HDevEngine stops due to a breakpoint or an error, HDevelop will automatically
switch to the corresponding thread.
The normal thread view in the dialog Execute . Thread View / Call Stack is not available, but threads can
be switched from this dialog by clicking the button Show Application Threads....
Please note that subthreads started with par_start cannot be debugged.
Click Execute . Stop Debugging to terminate the debugging session. The debug server will continue to run
and accept further connections as long as HDevEngine keeps running. In case the external application is currently
stopped at the moment the session is terminated, HDevelop will offer to resume the execution before disconnecting.
Otherwise, the thread executing the procedure calls will be frozen until you reconnect and continue running it.
The debugging session will also be terminated if the external application itself stops the debug server or if the
session is taken over by a subsequent connection.
9.9 Limitations
Further limitations of remote debugging from the HDevEngine side are listed in the Programmer’s Guide, sec-
tion 25.2.4 on page 205.
Code Export 303
Chapter 10
Code Export
The idea of code export or code generation is as follows: After developing a program according to the given
requirements it has to be translated into its final environment. For this, the program is transferred into another
programming language that can be compiled.
For C++ and C# developers there is a recommended approach for integrating HDevelop code into their own
projects. Instead of exporting an entire HDevelop program to C++ or C#, a library project can be exported,
which can be directly integrated into your own projects. The procedure can be called as simple as in HDevelop.
The library project export is described in section 10.1.
Code Export
Alternatively, HDevelop allows you to export an entire HDevelop program to the programming languages C++,
Visual Basic.NET, C#, and C by writing the corresponding code to a file. This approach is described in section 10.2
on page 307.
10.1.1 Requirements
10.1.1.1 C++
• The exported code makes use of some C++11 elements. These will not work with old compilers. The
minimum requirements for the target compilers are: VS 2013, GCC 4.8.6, clang 3.9, and ICC 16. Older
compilers might work but are not supported.
10.1.1.2 C#
Before you can export a library project, you need to create a procedure library containing all procedures that will
be required in your application. Alternatively, create a HDevelop program containing the required functionality in
local procedures. In the latter case, the main procedure will not be exported.
The declaration of public and private procedures defines the interface for your application. Public procedures
can be called from your application. Private procedures can only be called from procedures within the library or
HDevelop program.
Select File . Export Library Project to open the export dialog (see figure 10.1).
10.1.4 Using the Exported Library Project 305
Source
Input file
Select a procedure library to be exported. If Current Program is selected, the local procedures of the
current program will be exported.
If the selected file becomes unavailable (for example, because the procedure path settings have been changed
in the meantime), an error message is displayed.
Target Project Settings
Target Language
Specify the target language of the exported project (C++ or C#).
Encoding
Specify the text encoding of the exported project (either Native or UTF-8). Selecting Native will use the
current code page on Windows and UTF-8 on other platforms.
Currently, the other target project settings can not be modified. The library project will be exported to a
CMake-based project. The exported library will be accessible through a HDevEngine wrapper. Building the
exported project will result in a static library that can be linked to an application.
Destination
Project Name
Enter the project name. The project name specifies the name of a subdirectory of the project location, which
will contain all generated files. If the project name is not a valid file name, an error message is displayed.
Code Export
Project Location
Specify the destination directory of the exported project. Relative paths are allowed – the path will be relative
to the current working directory (see get_current_dir).
Namespace
Specify the namespace of the exported library project. The namespace must be valid for the target language.
Otherwise, an error message is displayed.
Export
Shows the final location of the library project based on the settings in Destination. After specifying the
project properties, click the button Export to generate the library project. If you export from HDevelop XL,
you will obtain an XL variant of the library project.
Previous exports at the same location will be silently overwritten.
!
The project directory will contain the following files:
• CMakeLists.txt
CMake control file.
• cmake (C++ only)
Directory containing CMake code for the resolution of HALCON and HDevEngine.
• res_project name
Directory containing the exported procedure library or HDevelop program.
• source
Directory containing the exported wrapper code and, for C++, a header file. In case of exporting Current
Program with no local procedure at all, a project directory will still be created. However, there will not be
any callable procedures/resources inside. If you experience this, try putting the relevant code of the main
procedure into a local procedure and repeat the library export.
The CMake project can easily be converted to a Visual Studio project via the Generate option. The following
example generates a project for Visual Studio 2015 (64-bit), but other generators are available as well.
306 Code Export
CMake
Move the exported project into a subdirectory of your CMake project and use it in your application project(s).
Assuming the exported library has the name MyLibrary, a minimal CMakeLists.txt file for an executable project
named ExportTest using the library look as follows:
project(ExportTest)
add_subdirectory(MyLibrary)
add_executable(ExportTest main.cpp)
target_link_libraries(ExportTest MyLibrary)
Visual Studio
• Add the library.
• Add the includes.
When linking the generated library project into the user application, both hdevenginecpp.lib and
halconcpp.lib (C++), or hdevenginedotnet.dll and halcondotnet.dll (C#) must be added to the project
as well. This is, because the HDevEngine symbols are not included in the exported library.
If you are working with CMake to configure and build your application, you can (re-)export directly into your
! application project file structure. However, be aware that HDevelop will silently overwrite any existing files in
the target location.
• Use the function SetResourcePath to set the directory containing the exported scripts.
• Include the exported header file.
• Call the procedures.
Note that the procedures only use parameters of type HObject, HTuple, HObjectVector, and HTupleVector.
HDevEngineCpp::HDevEngine().AddProcedurePath(...)
10.2 Exporting Entire HDevelop Programs 307
• Use the function ResourcePath to set the directory containing the exported scripts.
• Call the procedures, for example,
MyNamespace.MyLibrary.my_procedure()
Note that the procedures only use parameters of type HObject, HTuple, HObjectVector, and HTupleVector.
HalconDotNet::HDevEngine().AddProcedurePath(...)
Code Export
10.2 Exporting Entire HDevelop Programs
HDevelop allows you to export a developed HDevelop program to the programming languages C++, Visual Ba-
sic.NET, C#, and C. The following sections describe the general steps of program development using this feature
including some language-specific details of the code generation and optimization aspects for the following lan-
guages:
Because HDevelop does more than just execute a HALCON program, the behavior of an exported program will
differ in some points from its HDevelop counterpart. A prominent example is that in HDevelop, all results are
automatically displayed, while in the exported programs you have to insert the corresponding display operators
explicitly. Section 10.2.5 on page 314 describes these differences in more detail.
This section describes how to create a HALCON application in C++, starting from a program developed in HDe-
velop.
Program Export
To export the program, use the menu File . Export Program... and select the language (C++ - HALCON/C++).
The created file contains the HDevelop program as C++ source code.
For every HDevelop procedure except the main procedure, the exported file contains a C++ procedure with the
corresponding name.
308 Code Export
Iconic input and output parameters of a procedure are declared as HObject and HObject*, respectively, while
control input and output parameters are declared as HTuple and HTuple*, respectively.
All procedures are declared at the beginning of the file. The program body of the HDevelop main procedure is
contained in a procedure action(), which is called in the function main().
action() and main() can be excluded from compilation by inserting the instruction #define NO_EXPORT_MAIN
at the appropriate position in the application.
To exclude the main() procedure from compilation only, use the instruction #define NO_EXPORT_APP_MAIN.
This can be useful if you want to integrate exported HDevelop code into your application through specific proce-
dure interfaces. In that case, there is typically no need to export the main procedure, which was probably used only
for testing the functionality implemented in the corresponding ’real’ procedures.
Besides the program code, the file contains all necessary #include instructions. All local variables (iconic as well
as control) are declared in the corresponding procedures. Iconic variables belong to the class HObject and all
other variables belong to HTuple.
add_executable(test.cpp)
target_link_libraries(test HALCON::CppInt)
If you want to build using HALCON XL, set the option HALCON_XL to ON or 1 during the configuration step. For
this, use the following syntax:
To configure the build and create the application, run the following commands:
mkdir build
cd build
cmake %HALCONEXAMPLES%/cpp/console
cmake --build .
For more information, see the Programmer’s Guide, chapter 7 on page 51.
10.2.1.2 Optimization
Optimization might be necessary for variables of class HTuple. This kind of optimization can either be done in
HDevelop or in the generated C++ code. In most cases, optimization is not necessary if you program according to
the following rules.
• Using the tuple concatenation, it is more efficient to extend a tuple at the “right” side, like:
T := [T,New]
T := [New,T]
T = HTuple(New).TupleConcat(T);
T[i] := New
can directly be translated into the efficient and similar looking code
T[i] = New;
There are only two classes that are used: HTuple for control parameters and HObject for iconic data. There is no
need for other classes as long as the program has the same functionality as in HDevelop. When editing a generated
program you are free to use any of the classes of HALCON/C++ to extend the functionality.
Code Export
Besides the restrictions mentioned in this section and in section 10.2.5 on page 314, please also check the descrip-
tion of the HDevelop operators in section 6.1.6 on page 60.
Exception Handling
In HDevelop, every exception normally causes the program to stop and report an error message in a dialog window.
This might not be useful in C++. The standard way to handle this in C++ is by using the try-catch mechanism.
This allows you to access the reason for the exception and to continue accordingly. HDevelop supports exception
handling using try-catch blocks, which is exported transparently to C++. Therefore, it is the recommended
method of handling errors in HDevelop programs that are going to be exported to C++.
For HDevelop programs containing error handling using dev_set_check('~give_error') (or
set_check('~give_error')), the corresponding code is automatically included. Every operator call, for
which it is assumed that the HALCON error mechanism is turned off, is enclosed in a try block followed by a
catch block. The latter handles the exception and assigns the corresponding HALCON error number to the error
variable activated by dev_error_var or to a local error variable otherwise.
Please note that a call of dev_set_check('~give_error') has no influence on the operator call. The exception
will always be raised.
This section describes how to create a HALCON application in C#, starting from a program developed in HDe-
velop. HALCON can be used together with C# based on the .NET interface of HALCON. A detailed description
of this interface can be found in the Programmer’s Guide, part III on page 61.
Export
To export the program, use the menu File . Export Program... and select the language (C# - HALCON/.NET).
The created file with extension “.cs” contains the HDevelop program as C# source code.
310 Code Export
%HALCONEXAMPLES%\c#\HDevelopTemplate
This project contains a form with a display window (HWindowControl) and a button labeled Run. Add the file
generated by HDevelop to the project in the Solution Explorer (Add Existing Item). Now the project is ready
for execution: Run the project and then press the button Run on the form, which will call the exported code.
Additional information about using the template can be found in the Programmer’s Guide, section 12.3.1 on page
88.
If the program has been exported using the option Use Export Template, the file created by HDevelop contains
a subroutine with the corresponding name for every HDevelop procedure except the main procedure, which is
contained in the subroutine action().
Otherwise, the file is exported as a standalone application. Iconic input and output parameters of a procedure
are passed as HObject and out HObject, respectively, while control input and output parameters are passed as
HTuple and out HTuple, respectively. The subroutine RunHalcon() contains a call to the subroutine action()
and has a parameter Window, which is of type HTuple. This is the link to the window on the form to which
all output operations are passed. In addition, another subroutine is created with the name InitHalcon(). This
subroutine applies the same initializations that HDevelop performs.
Most of the variables, iconic as well as control, are declared locally inside the corresponding subroutines. Iconic
variables belong to the class HObject and control variables belong to HTuple.
Depending on the program, additional subroutines and variables are declared.
Stop
If the program has been exported using the option Use Export Template, the HDevelop operator stop is trans-
lated into a subroutine in C# that creates a message box. This message box causes the program to halt until the
button is pressed.
If the program has been exported using the option Use HALCON Windows, the operator stop will be ignored as
this export is mainly intended to compile and run standalone.
Used Classes
There are only four classes/types that are used: HTuple for control parameters and HObject for iconic data. In
addition, there is the class HWindowControl. It is used inside the project for the output window and a variable
of class HTuple directs the output to this window. Finally, the class HOperatorSet is used as a container for all
HALCON operators. There is no need for other classes as long as the program has the same functionality as in
HDevelop. When editing a generated program you are free to use any of the classes of HALCON/.NET to extend
the functionality.
Besides the restrictions mentioned in this section and in section 10.2.5 on page 314, please also check the descrip-
tion of the HDevelop operators in section 6.1.6 on page 60.
Variable Names
The export adds the prefix ho_ to all local iconic and hv_ to all local control variables, respectively, to avoid
collisions with reserved words.
10.2.3 Code Generation for Visual Basic.NET (HALCON/.NET) 311
Exception Handling
In HDevelop, every exception normally causes the program to stop and report an error message in a dialog window.
This might not be useful in C#. The standard way to handle this in C# is by using the try-catch mechanism. This
allows you to access the reason for the exception and to continue accordingly. For HDevelop programs contain-
ing error handling using dev_set_check('~give_error') (or set_check('~give_error')), the corresponding
code is automatically included. Every operator call, for which it is assumed that the HALCON error mechanism
is turned off, is enclosed in a try block followed by a catch block. The latter handles the exception and assigns
the corresponding HALCON error number to the error variable activated by dev_error_var or to a local error
variable otherwise.
Please note that a call of dev_set_check('~give_error') has no influence on the operator call. The exception
will always be raised. This is also true for messages like H_MSG_FAIL, which are not handled as exceptions in
C++, for example.
Memory Management
The .NET Framework’s runtime environment CLR (Common Language Runtime) has a mechanism called garbage
collector, which is used by the CLR to remove no longer needed .NET objects from memory. As mentioned earlier,
in the exported C# code every iconic object is represented by a .NET HObject object. From the garbage collector’s
point of view, a .NET HObject object is rather small. Thus, it might not be collected from memory although the
underlying iconic object (for example, an image) might in fact occupy a large portion of memory. To avoid memory
leaks caused by this effect, in the exported code every iconic object is disposed explicitly before it is assigned a
new value.
Code Export
10.2.3 Code Generation for Visual Basic.NET (HALCON/.NET)
This section describes how to create a HALCON application in Visual Basic.NET, starting from a program de-
veloped in HDevelop. HALCON can be used together with Visual Basic .NET based on the .NET interface of
HALCON. A detailed description of this interface can be found in the Programmer’s Guide, part III on page 61.
Export
To export the program, use the menu File . Export Program... and select the language (Visual Basic.NET
- HALCON/.NET). The created file with extension “.vb” contains the HDevelop program as Visual Basic.NET
source code.
%HALCONEXAMPLES%\vb.net\HDevelopTemplate
This project contains a form with a display window (HWindowControl) and a button labeled Run. Add the file
generated by HDevelop to the project in the Solution Explorer (Add Existing Item). Now the project is ready
for execution: Run the project and then press the button Run on the form, which will call the exported code.
Additional information about using the template can be found in the Programmer’s Guide, section 12.3.1 on page
88.
If the program has been exported using the option Use Export Template, the file created by HDevelop contains
a subroutine with the corresponding name for every HDevelop procedure except the main procedure, which is
contained in the subroutine action(). Otherwise, the file is exported as a standalone application. Iconic input and
312 Code Export
output parameters of a procedure are passed as ByVal HObject and ByRef HObject, respectively, while con-
trol input and output parameters are passed as ByVal HTuple and ByRef HTuple, respectively. The subroutine
RunHalcon() contains a call to the subroutine action() and has a parameter Window, which is of type HTuple.
This is the link to the window on the form to which all output operations are passed. In addition, another subroutine
is created with the name InitHalcon(). This subroutine applies the same initializations that HDevelop performs.
Most of the variables, iconic as well as control, are declared locally inside the corresponding subroutines. Iconic
variables belong to the class HObject and control variables belong to HTuple.
Depending on the program, additional subroutines and variables are declared.
Stop
If the program has been exported using the option Use Export Template, the HDevelop operator stop is trans-
lated into a subroutine in Visual Basic.NET that creates a message box. This message box causes the program to
halt until the button is pressed.
If the program has been exported using the option Use HALCON Windows, the operator stop will be ignored as
this export is mainly intended to compile and run standalone.
Exit
The HDevelop operator exit is translated into the Visual Basic.NET routine End. Because this routine has no
parameter, the parameters of exit are suppressed.
Used Classes
There are only four classes/types that are used: HTuple for control parameters and HObject for iconic data. In
addition, there is the class HWindowControl. It is used inside the project for the output window and a variable
of class HTuple directs the output to this window. Finally, the class HOperatorSet is used as a container for all
HALCON operators. There is no need for other classes as long as the program has the same functionality as in
HDevelop. When editing a generated program you are free to use any of the classes of HALCON/.NET to extend
the functionality.
Besides the restrictions mentioned in this section and in section 10.2.5 on page 314, please also check the descrip-
tion of the HDevelop operators in section 6.1.6 on page 60.
Variable Names
In contrast to C, C++, or HDevelop, Visual Basic.NET has many reserved words. Thus, the export adds the prefix
ho_ to all iconic and hv_ to all control variables, respectively, to avoid collisions with these reserved words. See
also section 10.2.5.3 on page 315 about case sensitivity.
Exception Handling
In HDevelop, every exception normally causes the program to stop and report an error message in a dialog win-
dow. This might not be useful in Visual Basic.NET. The standard way to handle this in Visual Basic.NET is
by using the Try-Catch mechanism. This allows you to access the reason for the exception and to continue
accordingly. For HDevelop programs containing error handling using dev_set_check('~give_error') (or
set_check('~give_error')), the corresponding code is automatically included. Every operator call, for which
it is assumed that the HALCON error mechanism is turned off, is enclosed in a Try block followed by a Catch
block. The latter handles the exception and assigns the corresponding HALCON error number to the error variable
activated by dev_error_var or to a local error variable otherwise.
Please note that a call of dev_set_check('~give_error') has no influence on the operator call. The exception
will always be raised. This is also true for messages like H_MSG_FAIL, which are not handled as exceptions in
C++, for example.
10.2.4 Code Generation for C 313
Memory Management
The .NET Framework’s runtime environment CLR (Common Language Runtime) has a mechanism called garbage
collector, which is used by the CLR to remove no longer needed .NET objects from memory. As mentioned earlier,
in the exported Visual Basic.NET code every iconic object is represented by a .NET HObject object. From the
garbage collector’s point of view, a .NET HObject object is rather small. Thus, it might not be collected from
memory although the underlying iconic object (for example, an image) might in fact occupy a large portion of
memory. To avoid memory leaks caused by this effect, in the exported code every iconic object is deleted explicitly
before it is assigned a new value.
Parallel Execution
If the program to be exported includes the parallel execution of procedure or operator calls using par_start (see
section 8.11 on page 283), the exported code requires Visual Studio 2010 or higher.
This section describes how to create a HALCON application in C, starting from a program developed in HDevelop.
Program Export
Code Export
The first step is to export the program using the menu File . Export Program.... Here, select the language
(C - HALCON/C) and save it to file. A file will be created that contains the HDevelop program as C source code.
For every HDevelop procedure except the main procedure, the exported file contains a C procedure with the cor-
responding name. Iconic input and output parameters of a procedure are declared as Hobject and Hobject*,
respectively, while control input and output parameters are declared as Htuple and Htuple*, respectively. All
procedures are declared at the beginning of the file. The program body of the HDevelop main procedure is con-
tained in a procedure action() which is called in function main(). action() and main() can be excluded from
compilation by inserting the instruction #define NO_EXPORT_MAIN at the appropriate position in the application.
Using the instruction #define NO_EXPORT_APP_MAIN only the main() procedure is excluded from compilation.
This can be useful if you want to integrate exported HDevelop code into your application through specific proce-
dure interfaces. In that case, there is typically no need to export the main procedure, which was probably used only
for testing the functionality implemented in the corresponding ’real’ procedures.
Besides the program code, the file contains all necessary #include instructions. All local variables (iconic as well
as control) are declared in the corresponding procedures. Iconic variables belong to the class Hobject and all
other variables belong to Htuple.
Please note that in the current version the generated C code is not optimized for readability. It is output such that
it always produces identical results as the HDevelop code.
add_executable(test test.c)
target_link_libraries(test HALCON::CInt)
If you want to build using HALCON XL, set the option HALCON_XL to ON or 1.
To configure the build and create the application, run the following commands:
314 Code Export
mkdir build
cd build
cmake %HALCONEXAMPLES%/c
cmake --build .
For more details see the Programmer’s Guide, chapter 19 on page 127.
In the following, general differences in the behavior of an HDevelop program and its exported versions are de-
scribed.
It is possible to embed arbitrary code into HDevelop programs. This code is ignored inside HDevelop. When you
export the program to a programming language, the embedded code is passed through verbatim.
Program lines starting with # as the first character mark an arbitrary code line. The marker and the first space
character following it are discarded when exporting the program. For example, the line
in Visual Basic. The # may be followed by other special characters that further specify where the code block will
be placed upon exporting. For example, the line
#define NO_EXPORT_APP_MAIN
at the very beginning of the exported program. Code lines in this format are collected from the main procedure
first, followed by #^^ lines in other procedures.
The recognized special markers are summarized in table 10.1.
If you are using the operator window to enter arbitrary code lines, you will have to select the special operator
export_def. Its first parameter specifies the destination of the exported code line (see the last column of table 10.1
for reference). The second parameter is the code line itself. When you submit the operator to the program window,
the operator call will be converted to the special prefix characters for better readability.
10.2.5 General Aspects of Code Generation 315
10.2.5.2 Assignment
In HDevelop each time a new value is assigned to a variable its old contents are removed automatically, independent
of the type of the variable. In the exported code, this is also the case for iconic objects (HALCON/C++: Hobject,
HALCON/.NET: HObject) and for the class HTuple (HALCON/C++, HALCON/.NET). Because C does not
provide destructors, the generated C code calls the operators clear_obj and destroy_tuple to remove the
content of iconic output parameters (Hobject) and control output parameters (Htuple) before each operator call.
Memory issues regarding iconic objects in HALCON/.NET are described in section 10.2.3.3 (Visual Basic.NET)
and section 10.2.2.3 (C#).
Variable names in HDevelop are case-sensitive, x and X are distinct variable names in HDevelop programs. If
you export such a program to a case-insensitive target language, for example Visual Basic.NET, the development
environment will complain about multiple declarations. Either plan ahead and avoid these variable names, or use
the Find/Replace dialog to substitute conflicting variable names before exporting your program.
As described for the different programming languages, HDevelop procedures are exported automatically to pro-
cedures or subroutines of the selected programming language. This does not hold for the protected procedures
described in section 5.6 on page 47. These procedures are protected by a password so that they cannot be viewed
Code Export
and modified by unauthorized users. Thus, as long as they are locked by the password, they can not be exported to
any programming language.
You should know that HDevelop performs some changes of system parameters of HALCON by calling the operator
set_system (see the reference manual). This might cause the exported program not to produce identical output. If
such a problem arises, you may query the system parameters by means of get_system in HDevelop after or while
running the original HDevelop version of the program. Depending to the problem, you can now modify relevant
parameters by explicitly calling the operator set_system in the exported program.
HALCON provides a functionality that emulates the behavior of HDevelop graphics windows for HALCON win-
dows. This HALCON window stack is accessible via class methods and functions in the HALCON interfaces,
and code exported from HDevelop uses this functionality when opening, closing, setting, or accessing the active
window. The HALCON window stack mechanism is threadsafe and can be shared among threads. Though, in
a multithreaded application the user has to take care when switching the active window in different threads, as
setting it in one thread will also set it for other threads.
For the .NET code export it is optional whether to export HDevelop programs as code using the HDevelop export
example templates or as code using the previously described HALCON window stack functionality when doing
graphics windows output. Additionally, in the latter case the exported code contains a main function and thus is
usable as a standalone application. The HDevelop Export dialog allows you to select the corresponding option.
The graphics windows of HDevelop and the basic windows of the HALCON libraries
• Multiple windows
If you use the operator dev_open_window to open multiple graphics windows in HDevelop, these calls will
be converted into corresponding calls of open_window only if the export option Use HALCON Windows is
selected.
In the export of Visual Basic.NET, and C# programs using the option Use Export Template, all window
operations are suppressed, because the exported code is intended to work together with the corresponding
template. If you want to use more than one window in programs exported in this mode, you have to modify
the code and project manually.
Note that the export of programs containing multiple windows using the option Use HALCON Windows might
be incorrect if the active graphics window was changed using the mouse during program execution. It is
recommended to use the operator dev_set_window explicitly to achieve the same functionality.
• Window size
In exported Visual Basic.NET, and C# programs, the size of the window on the form is predefined (512 ×
512); thus, it will normally not fit your image size. Therefore, you must adapt the size interactively or by
using the properties of the window.
• Displaying results
Normally, the result of every operator is displayed in the graphics window of HDevelop. This is not the case
when using an exported program. It behaves like the HDevelop program running with the option: “update
window = off”. We recommend inserting the operator dev_display in the HDevelop program at each point
where you want to display data. This will not change the behavior of the HDevelop program but result in the
appropriate call in the exported code.
When generating code using the option Use HALCON Windows, close the default graphics window (using
dev_close_window) and open a new one (using dev_open_window) before the first call of dev_display
in order to assure a correct export.
• Displaying images
In HDevelop, images are automatically scaled to fit the current window size. This is not the case in exported
programs. For example, if you load and display two images of different size, the second one will appear
clipped if it is larger than the first image or filled up with black areas if it is smaller. For a correct display,
you must use the operator dev_set_part before displaying an image with dev_display as follows:
In this example, Image is the image variable, ImageHeight and ImageWidth denote its size. You can query
the size of an image with the operator get_image_size. Please consult the HALCON Reference Manuals
for more details.
Note that the operator dev_set_part (and its HALCON library equivalent set_part) is more commonly
used for displaying (and thereby zooming) parts of images. By calling it with the full size of an image as
shown above, you assure that the image exactly fits the window.
• Changing display parameters
If you change the way how results are displayed (color, line width, etc.) in HDevelop interactively via the
menu Visualization, these changes will not be incorporated in the exported program. We recommend
inserting the corresponding Develop operators (for example, dev_set_color or dev_set_line_width)
in the HDevelop program explicitly. This will result in the appropriate call (set_color, set_line_width,
etc.) in the exported code.
HDevelop and exported code can behave differently regarding the values of output parameters of procedures in
case an output parameter is not assigned a value inside the procedure. This can happen if the procedure returns
with an exception before setting an output parameter or if the parameter was not set on purpose, in which case
the value of the output parameter is not defined and thus HDevelop and the different code exports may behave
differently. The main reason why the code exports are not adapted to behave like HDevelop is that that would
imply a deterioration of runtime efficiency in the exported code.
10.2.5 General Aspects of Code Generation 317
When accessing uninitialized elements of a tuple or vector with HDevelop, an exception is raised. Due to per-
formance reasons, there is no corresponding check in the HALCON language interfaces. Instead of raising an
exception, an element of the default type is returned.
a := [0, 1]
b := a[2]
For example, this code snippet raises an exception in HDevelop, but might return an arbitrary value for b in one of
the language interfaces. Thus, this coding style must be avoided.
Code Export
318 Code Export
Glossary 319
Appendix A
Glossary
Boolean is the type name for the truth values true and false as well as for the related boolean expressions.
Body A body is part of a conditional instruction (if) or a loop (while or for) and consists of a sequence of
operator calls. If you consider the for-loop, for instance, all operator calls, that are located between for
and endfor form the body.
Control data Control data can be either numbers (↑integer and ↑real), character strings (↑string) and truth
values (boolean). This data can be used as atomic values (i.e., single values) or as ↑tuples (i.e., arrays of
values).
Empty region An empty ↑region contains no points at all, i.e., its area is zero.
Graphics window A graphics window is used in ↑HDevelop for displaying, e.g., ↑images, ↑regions, and ↑XLD.
HDevelop is an interactive program for the creation of HALCON applications.
Iconic data are image data, i.e., image arrays and data, which are described by coordinates and are derived from
Glossary
image arrays, e.g., ↑regions, ↑images and ↑XLD.
Image An image consists of one or more (multi-channel image) image arrays and a ↑region as the definition
domain. All image arrays have the same dimension, but they can be of different pixel types. The size of the
↑region is smaller or equal than the size of the image arrays. The ↑region determines all image points that
should be processed.
Iconic object Generic implementation of ↑iconic data in HALCON.
integer is the type name for integer numbers.
Operator data base The operator data base contains information about the HALCON operators. It is loaded at
runtime from the binary files in %HALCONROOT%\help.
Program Window In HDevelop the program window contains the program. It is used to edit (copy, delete, and
paste lines) and to run or debug the program.
Operator window In the operator window of HDevelop the parameters of the selected operators can be entered
or modified.
Real is the type name for floating point numbers. They are implemented using the C-type double (8 bytes).
Region A region is a set of image points without gray values. A region can be imagined as a binary image (mask).
Regions are implemented using runlength encoding. The region size is not limited to the image size (see
also set_system(’clip_region’,’true’/’false’) in the HALCON reference manual).
String is the type name for character strings. A string starts and ends with a single quote; in between any character
can be used except single quote. The empty string consists of two consecutive single quotes.
Tuple A tuple is an ordered multivalue set. In case of ↑control data a tuple can consist of a large number of
items with different data types. The term tuple is also used in conjunction with ↑iconic objects, if it is to be
emphasized that several ↑iconic objects will be used.
320 Glossary
Type ↑iconic variables can be assigned with data items of type ↑image, ↑region, and ↑XLD. The types of ↑control
data items can be one of ↑integer, ↑real, ↑boolean, or ↑string.
Variable window In HDevelop the variable window manages the ↑control and ↑iconic data.
XLD is the short term for eXtended Line Description. It is used as a superclass for contours, polygons, and lines.
Color Names 321
Appendix B
Color Names
Color Names
slate blue sample #6A5ACD
spring green sample #00FF7F
orange red sample #FF4500
dark olive green sample #556B2F
pink sample #FFC0CB
cadet blue sample #5F9EA0
goldenrod sample #DAA520
orange sample #FFA500
gold sample #FFD700
forest green sample #228B22
cornflower blue sample #6495ED
navy sample #000080
turquoise sample #40E0D0
dark slate blue sample #483D8B
light blue sample #ADD8E6
322 Color Names
Command Line
Appendix C
C.1 HDevelop
HDevelop understands many command line switches when started from a console or terminal window. Run the
following command to get a list of all supported command line switches:
hdevelop -h
FILE can be an HDevelop program, an external procedure, a library file, or an image file. The default action is to
load the given file for editing. If an image file is given, it is preselected in the dialog Read Image... (page 24)
(same behavior as dragging an image file onto a running instance of HDevelop).
Examples
• Load the HDevelop program example.hdev and run it immediately:
hdevelop -run example.hdev
• Use the procedures and libraries from C:/projects in the next session:
hdevelop -external_proc_path:C:/projects"
Specified path names are put first in the list of directories that are searched for external procedures and
libraries. Separate multiple directories by the system-dependent separator (“;” on Windows and “:” on
Linux). This is a temporary setting for the current session only. See also section 5.4 on page 45.
• Open the protected program secret.hdev for editing (password: crypt):
hdevelop -unlock:crypt secret.hdev
hdevelop [options]
HDevelop options:
<program>.{hdev,dev}
load the program for editing, converting, or running
{<procedure_file>,<procedure_library_file>}
load the procedure or procedure library for editing
-{protect,unprotect,unlock}:<password>
modify the protection state of an HDevelop program or
procedure
-load_recent_prog load the recent program irrespective of the appropriate
option that was selected in the preferences dialog
-load_no_prog do not load the recent program irrespective of the
appropriate option that was selected in the preferences
dialog
-run start execution of the passed program
<image_file> load an image file with read_image
-help show this help info in a message box
--help show this help information on the console
-version show version information in a message box
--version show version information on the console
-external_proc_path:<external procedure path(s)>
an external procedure path may point either to a directory
or a procedure library file
multiple external procedure paths are separated by semicolons
on windows systems, or by colons on all other systems
-convert <source> <destination> [<options>]
convert an HDevelop program or procedure into a
file of the specified type
(not possible combinations:
convert dev, hdev, hdpl into dvp, hdvp
convert dev, hdev, dvp, hdvp into hdpl)
source: <src_file>.{hdev,dev}
<src_file>.{hdvp,dvp}
<src_file>.{hdpl}
destination: dest_file.<type>
<type> write to <src_file>.<type>
- <type> write to stdout
type: hdev,dev HDevelop program
hdvp,dvp HDevelop procedure
hdpl HDevelop procedure library
c,cpp C, C++
C.1 HDevelop 325
Command Line
options: [-external_procs_only_interfaces]
export only the interface(s) of the procedures
(the declarations) without the bodies
[-no_export_of_referred_procs]
export only the passed program or procedure but
do not include any referred external procedures
[-no_msg_box]
suppress error messages
[-reset_free_text]
reset free text formating
[-delete_local_procs]
[-delete_unused_local_procs]
[-no_use_hdevelop_template]
[-encoding native]
exported string constants are encoded in native
local-8-bit encoding instead of UTF-8
[-encoding UTF-8]
exported string constants are encoded in UTF-8 (default)
Please note that non-ASCII characters in string constants
are exported as octal codes in order to guarantee that
the strings are correctly created on all systems,
independent on any compiler settings
-reset_preferences
reset all persistent settings from previous sessions
-add_preferences <file>
start HDevelop with additional preferences
from a file
(replaces -preferences <file> that became deprecated)
-load_preferences <file>
reset all persistent settings and
start HDevelop with the preferences from a file
-use_preferences <file>
start HDevelop with the preferences from file
and store all modified preferences to file
-override_stop <time>
override stop() operator with wait_seconds(time)
-override_wait <time>
replace time of wait_seconds() operator with <time>
Namespace of project
(must be a valid identifier in the target language)
-language <language>
Set source code language. Currently supported:
cpp
cs
GUI options:
-style[=] <style>
sets the application GUI style. Possible values are:
Windows Motif CDE Plastique Cleanlooks
X11 options:
-display <display>
sets the X display (default is $DISPLAY).
-geometry <geometry>
sets the client geometry of the first window that is shown.
-graphicssystem {native|raster|opengl}
sets the graphics backend used (default: native).
-{fn|font} <font>
defines the application font. The font should be specified
using an X logical font description.
-{bg|background} <color>
sets the default background color and an application palette
(light and dark shades are calculated).
-{fg|foreground} <color>
sets the default foreground color.
-{btn|button} <color>
sets the default button color.
-name <name> sets the application name.
-title <app_title> sets the application title.
-visual TrueColor forces the application to use a TrueColor visual on an
8-bit display.
-ncols <count> limits the number of colors allocated in the color cube on
an 8-bit display if the application is using the
QApplication::ManyColor color specification. If count is 216
then a 6x6x6 color cube is used (i.e., 6 levels of red,
6 of green, and 6 of blue); for other values, a cube
approximately proportional to a 2x3x1cube is used.
-cmap causes the application to install a private color map on an
8-bit display.
C.2 Hrun
hrun --help
Appendix D
Keyboard Shortcuts
Keyboard Shortcuts
Keyboard Shortcuts
Activate F3
Deactivate F4
Find: Ctrl+F
Find Again Ctrl+G
Reset Program Execution F2
Reset Procedure Execution Shift+F2
Run F5
Run To Insert Cursor Shift+F5
Step Over F6
Step Forward Shift+F6
Step Into F7
Step Out F8
Abort Procedure Execution Shift+F8
Stop F9
Stop After Procedure Shift+F9
Set Breakpoint F10
Activate/Deactivate Breakpoint Shift+F10
Next Bookmark F11
Previous Bookmark Shift+F11
Set Bookmark Ctrl+F11
Activate Profiling Ctrl+Shift+F,F Ctrl+Shift+F,Ctrl+Shift+F
Number of Calls Ctrl+Shift+F,N Ctrl+Shift+F,Ctrl+Shift+N
Total Execution Time Ctrl+Shift+F,L Ctrl+Shift+F,Ctrl+Shift+L
Average Execution Time Ctrl+Shift+F,A Ctrl+Shift+F,Ctrl+Shift+A
Minimum Execution Time Ctrl+Shift+F,I Ctrl+Shift+F,Ctrl+Shift+I
Maximum Execution Time Ctrl+Shift+F,X Ctrl+Shift+F,Ctrl+Shift+X
Last Execution Time Ctrl+Shift+F,M Ctrl+Shift+F,Ctrl+Shift+M
Execution Time Ctrl+Shift+F,E Ctrl+Shift+F,Ctrl+Shift+E
Operator Time Ctrl+Shift+F,O Ctrl+Shift+F,Ctrl+Shift+O
Time Ctrl+Shift+F,T Ctrl+Shift+F,Ctrl+Shift+T
Percentage Ctrl+Shift+F,P Ctrl+Shift+F,Ctrl+Shift+P
Reset Profiler Ctrl+Shift+F,R Ctrl+Shift+F,Ctrl+Shift+R
Show Runtime Statistics Ctrl+Shift+F,S Ctrl+Shift+F,Ctrl+Shift+S
Activate/Deactivate display Ctrl+Shift+F,V Ctrl+Shift+F,Ctrl+Shift+V
Activate only selected Ctrl+Shift+F,Y Ctrl+Shift+F,Ctrl+Shift+Y
330 Keyboard Shortcuts
Keyboard Shortcuts
Full Screen Ctrl+Shift+W,F Ctrl+Shift+W,Ctrl+Shift+F
Next Window Ctrl+Tab Ctrl+>
Previous Window Ctrl+Shift+Tab Ctrl+<
Help F1
Context Help Shift+F1
Start Dialog Ctrl+Shift+H,S Ctrl+Shift+H,Ctrl+Shift+S
About Ctrl+Shift+H,A Ctrl+Shift+H,Ctrl+Shift+A
Keywords Ctrl+Shift+H,K Ctrl+Shift+H,Ctrl+Shift+K
HALCON News (WWW) Ctrl+Shift+H,W Ctrl+Shift+H,Ctrl+Shift+W
Search Documentation Ctrl+Shift+H,F Ctrl+Shift+H,Ctrl+Shift+F
HDevelop Language Ctrl+Shift+H,L Ctrl+Shift+H,Ctrl+Shift+L
HALCON Reference Ctrl+Shift+H,R Ctrl+Shift+H,Ctrl+Shift+R
HDevelop User’s Guide Ctrl+Shift+H,U Ctrl+Shift+H,Ctrl+Shift+U
Keyboard Shortcuts
Select a Procedure Alt+Up
Move tab to the left Ctrl+Alt+Shift+Left
Move tab to the right Ctrl+Alt+Shift+Right
Incremment text size Ctrl++
Decremment text size Ctrl+-
Set Breakpoint F10
Activate/Deactivate Breakpoint Shift+F10
Set Bookmark Ctrl+F11
Back in History Alt+Shift+Left
Forward in History Alt+Shift+Right
New Tab Alt+Ins
List Open Tabs Alt+Down
Index
* (asterisk) assistant
external procedure, 53 Close Dialog, 179
in window title, 32 Delete Generated Code Lines, 179
.NET, 309, 311 Exit Assistant, 179
.avi, 27 image acquisition, 25
.dev, 44 Insert Code for Selection, 179
.dvp, 44 Load Assistant Settings..., 179
.hdev, 44 Release Generated Code Lines, 179
Index
.hdpl, 45 Save Current Assistant Settings...,
.hdvp, 44 179
.seq, 27 Show Code Preview, 179
#, 314 assistant settings, display, 204
#$, 314 assistant settings, load, 199, 204
#$$, 314 assistant settings, save, 199, 204
#ˆ, 314 Assistants
#ˆˆ, 314 menu, 62
$, 260 Attach To Process..., 56
Attention, 143
Abort Procedure Execution, 56 Auto, 162
About, 64 Auto Disconnect, 183
Acquisition Auto Indent, 134
menu (Image Acquisition Assistant), 183 Auto-detect Interfaces, 180
Snap, 180 Autocompletion, 113
Acquisition menu autocompletion, 130
Live, 180 AVI, 27
Acquisition Mode, 183
actionUpdateWin, 58 Back, 85
Activate, 54, 64, 134 Back in History, 127
Activate Profiler, 56, 64 basic features, 237
Activate/Deactivate Breakpoint, 56, 134 Bit Depth, 181
Activate/Deactivate Breakpoint on Bookmark, 85
Variable, 135, 164, 166 Bookmarks, 155
Adapt program, 139 boolean, 248, 319
Adaptive Range, 98, 173 operations, 267
Add to User Tab, 164, 166 breakpoint, 17
Add Variable, 165, 167 activate/deactivate, 56
Add Watch, 135 clear all, 56
Advanced, 143 manage, 56
advanced autocompletion, 130 on variable, 162
advanced fuzzy features, 227 set, 56
advanced measuring tasks, 224, 234 Bring Main Window To Front, 63
advanced model parameters, 206 Bring Others To Front, 63
advanced search parameters, 216 Browse HDevelop Example Programs..., 53,
advanced training parameters, 238 64, 69
alignment, 228
All, 162 C, 11, 313
Alternatives, 61 C export, 313
Always Find, 206, 214 C++, 11
Angle Extent, 210 compile and link (Linux), 308, 313
Angle Step, 212 compile and link (Windows), 308, 313
Apply Changes Immediately, 58 C++ export, 307
336 Index
Index
Display Content:, 164 example programs, 53
Display Grid, 80 exception
Display Image, 182 handling, 283
Display Image Pyramid, 209 throw directly, 123
Display Image Pyramid Level, 204 Exception handling, 309, 311, 312
display parameters, 194, 222, 241 executable expression, 272
Display Selected Test Image, 206, 214 Execute
Draw, 56, 103 Activate/Deactivate Breakpoint, 56
Draw Axis-aligned Rectangle, 243 Attach To Process..., 56
Draw Circular Arc, 229 menu, 55
Draw Line, 229 Stop Debugging, 56
Draw Rotated Rectangle, 243 Execute Current Line, 128
Draw1, 159 Execute menu
Duplicate..., 59 Abort Procedure Execution, 56
Activate Profiler, 56
EasyParams, 29 Clear All Breakpoints, 56
edge data, 223 Clear Breakpoint, 56
edge selection, 222 Manage Breakpoints, 56
Edit menu, 54 Reset Procedure Execution, 56
Activate, 54 Reset Profiler, 56
Copy, 54 Reset Program Execution, 56, 251
Cut, 54 Run, 56
Deactivate, 54 Run To Insert Cursor, 56
Delete, 54 Set Breakpoint, 56
Find Again, 54 Show Runtime Statistics, 56
Find/Replace..., 54 Step Forward, 56
Invalid Lines, 54 Step Into, 56
Manage Bookmarks, 54 Step Out, 56
Next Bookmark, 54 Step Over, 56
Paste, 54 Stop, 56
Preferences, 54 Thread View / Call Stack, 56
Previous Bookmark, 54 Exit Assistant, 179, 199, 204, 229, 243
Redo, 54 expected gray value range, 224
Set/Clear Bookmark, 54 exponential
Undo, 54 functions, 268
Edit Procedure, 59 Export, 112
Edit Procedure Interface, 59, 128 Export Library Project..., 303
else, 277 Export Program..., 69
Emergency backup, 297 Export...
Enable the context menu in the Graphics UTF-8 encoding, 71
window, 126 external procedure, modified, 53
Enable the mouse wheel in the Graphics extract edges, 221
window, 126
338 Index
Index
image properties Load Image, 228, 242
status bar, 65 Load Model, 204
image pyramid, display, 209 Load Model Image, 204
Image Size, 58 Load OCR Classifier, 242
image source, 188, 220, 233 Load ROI from File., 229, 243
Import, 112 Load Test Images, 206
In Single Step Mode, 58 Load Training File, 242
Increase Zoom Factor, 85 local variables, 251
Increasing Range, 98, 173 Locate page, 85
Indent Size, 113 Logarithmic Scale, 98, 173
Input Window, 101 look-up table, 58
Insert All As Local, 59 loop
Insert Code, 183, 199, 207, 229, 243 body, 319
Insert Code for Selection, 179 Loop Counter, 183
Insert Code..., 58 Lut, 58
insert cursor, 17
Insert dev_display() into program, 164 main procedure, 307, 308, 310, 311, 313
Insert operator into program, 85 main window, 51
Insert Plot Code for Graphics Window, 99, Make All External, 59
174 Manage Breakpoints, 56, 135
Insert Used As Local, 59 Manage Passwords, 119
InsertCode, 102 Manage Procedures, 59
Inspect, 166, 207 Max Deformation, 216
Image Acquisition Assistant, 182 Max. Column Scale, 210
Inspect as Handle, 166 Max. Row Scale, 210
Inspect as Tuple, 166, 174 Maximum Number of Matches, 215
inspection, 235 Maximum Overlap, 216
integer, 248 measure task, setup, 220
Interface Library, 181 Measuring, 229
Invalid Lines, 153 menu, 52
Acquisition (Image Acquisition Assistant),
JIT compilation, 48 183
join, 264 Assistants, 62
Jump to Link, 85 Edit, 54
Execute, 55
Keep dialog open, 71 File, 52
keyboard Help, 64
menu access, 52 Operators, 60
shortcuts, 327 Procedures, 59
usage, 13 Suggestions, 61
Keywords, 61 Visualization, 56
Window, 63
LANG, 114 messages
340 Index
Index
Predecessors, 61 Release Generated Code Lines, 179, 200, 207,
Preferences, 112 229, 244
preferences remote debugging, 299
export, 112 Remove, 139, 174
import, 112 Remove All Test Images, 206
Preview Code, 224, 242 Remove from User Tab, 165, 167
Previous, 85 Remove Test Image, 206
Previous Bookmark, 54 repeat, 279
Print..., 85, 111, 135 replace
procedure, 307, 310, 311, 313 Find/Replace..., 54, 72, 154
procedure compilation, 48 Replace selected program lines, 139
procedures, 43 reserved words, 276
Procedures menu Reset, 112, 138
Create New Procedure, 59 reset, 236
Delete All Unused Local, 59 graphics window, 58
Delete Current, 59 Reset All, 27, 181
Edit Procedure, 59 Reset Bounds, 98, 99, 172–174
Edit Procedure Interface, 59 Reset Parameters, 58
Insert All As Local, 59 Reset Procedure Execution, 56, 64
Insert Used As Local, 59 Reset Profiler, 56
Make All External, 59 Reset Program Execution, 56, 64
Manage Procedures, 59 Reset to normal size, 85
process features, 223 Resolution, 181
Profiler Display, 56 Restore Default Layout Use, 63
program counter, 17 Restrictions, 309, 310, 312
Program Protection State:, 119 results, 241
Program Window, 17, 319 results tab, 222, 240
program window, 127 ROI Type, 158
Properties, 111 row, 65
Pyramid Levels, 210 Row Scale Step, 212
pyramid, display, 209 Run, 56, 64
Run To Insert Cursor, 56
quality issues, 191 Run Until Here, 133
Quick Navigation, 152 runtime
quick setup, 232 status bar, 65
Quit, 53
Save, 53, 64, 164, 166
Read Image..., 24 Save All, 53, 64
real, 248 Save Current Assistant Settings, 199, 204,
Recent Programs, 53 228, 243
recognition rate, determine, 218 Save Current Assistant Settings..., 179
recognition speed, optimize, 217 Save Model, 204
Record Interactions, 57 Save Procedure As..., 53
342 Index
Index
Clear Graphics Window, 58
Themes, 115 Close Graphics Window, 58
Thickness, 174 Color, 56
Thread View / Call Stack, 56, 64, 160 Colored, 58
throw, 282 Display, 58
Tilt, 212 Draw, 58
Timeout, 216 Feature Histogram, 58, 104
To Object / To Tuple, 139 Feature Inspection, 58, 105
To Vector, 139 Gray Histogram, 58, 99
Tools, 58 Image Size, 58
Train Now, 243 Insert Code..., 58
training, 237 Line Profile, 58, 107
Trigger, 181 Line Width, 58
trigonometric Lut, 58
functions, 267 New Zoom Window, 58
true, 319 OCR Training File Browser, 58, 86
try, 282 Open Graphics Window..., 58
tuple, 250, 319 Paint, 58
concatenation, 255, 256 Position Precision, 58
Type, 135 Reset Parameters, 58
type, 320 Save Window..., 58
boolean, 319 Set Parameters, 58
integer, 319 Shape, 58
real, 319 Update Window, 58
string, 319 Window Size, 58
Zoom Window, 58, 177
Undo, 54, 64 Volatile, 182
Ungroup, 174
Unicode, 71 Warning, 144
Unit, 159 while
until, 279 loop, 279
Update Graphics Window, 126 while, 319
Update Image, 181 Window
Update Program Counter, 125, 134 Cascade Floating Windows, 63
Update Variables, 125, 164, 167 menu, 63
Update Window, 58 Open Graphics Window, 63
Use Model, 206 Open Operator Window, 63
User, 162 Open Output Console, 63
User-defined Range, 98, 173 Open Program Window, 63
UTF-8 encoding, 71, 122 Open Quick Navigation, 63
Open Variable Window, 63
variable names, 224, 242 Window Size, 58
variable types, 251 Windows, 308, 313
344 Index
X, 181
XLD, 320
colors, 58
line width, 58
Y, 181
Zoom, 177
Zoom in, 128
Zoom In Mode, 99, 174
Zoom out, 128
Zoom Out Mode, 99, 174
Zoom To Selection, 102, 108
Zoom Window, 58, 64, 177