Computer Vision System Toolbox

Getting Started Guide

R2016a

How to Contact MathWorks

Latest news:

www.mathworks.com

Sales and services:

www.mathworks.com/sales_and_services

User community:

www.mathworks.com/matlabcentral

Technical support:

www.mathworks.com/support/contact_us

Phone:

508-647-7000

The MathWorks, Inc.

3 Apple Hill Drive
Natick, MA 01760-2098
Computer Vision System Toolbox Getting Started Guide
COPYRIGHT 20002016 by The MathWorks, Inc.
The software described in this document is furnished under a license agreement. The software may be used
or copied only under the terms of the license agreement. No part of this manual may be photocopied or
reproduced in any form without prior written consent from The MathWorks, Inc.
FEDERAL ACQUISITION: This provision applies to all acquisitions of the Program and Documentation
by, for, or through the federal government of the United States. By accepting delivery of the Program
or Documentation, the government hereby agrees that this software or documentation qualifies as
commercial computer software or commercial computer software documentation as such terms are used
or defined in FAR 12.212, DFARS Part 227.72, and DFARS 252.227-7014. Accordingly, the terms and
conditions of this Agreement and only those rights specified in this Agreement, shall pertain to and
govern the use, modification, reproduction, release, performance, display, and disclosure of the Program
and Documentation by the federal government (or other entity acquiring for or through the federal
government) and shall supersede any conflicting contractual terms or conditions. If this License fails
to meet the government's needs or is inconsistent in any respect with federal procurement law, the
government agrees to return the Program and Documentation, unused, to The MathWorks, Inc.

Trademarks

MATLAB and Simulink are registered trademarks of The MathWorks, Inc. See
www.mathworks.com/trademarks for a list of additional trademarks. Other product or brand
names may be trademarks or registered trademarks of their respective holders.
Patents

MathWorks products are protected by one or more U.S. patents. Please see
www.mathworks.com/patents for more information.

Revision History

April 2011
September 2011
March 2012
September 2012
March 2013
September 2013
March 2014
October 2014
March 2015
September 2015
March 2016

Online only
Online only
Online only
Online only
Online only
Online only
Online only
Online only
Online only
Online only
Online only

Revised for Version 4.0 (Release 2011a)

Revised for Version 4.1 (Release 2011b)
Revised for Version 5.0 (Release 2012a)
Revised for Version 5.1 (R2012b)
Revised for Version 5.2 (R2013a)
Revised for Version 5.3 (R2013b)
Revised for Version 6.0 (R2014a)
Revised for Version 6.1 (R2014b)
Revised for Version 6.2 (Release R2015a)
Revised for Version 7.0 (Release R2015b)
Revised for Version 7.1 (Release R2016a)

Contents

Product Overview
Computer Vision System Toolbox Product Description . . . .
Key Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1-2
1-2

Computer Vision Algorithms and Video Processing

Computer Vision Capabilities . . . . . . . . . . . . . . . . . . . . . . . . .

2-2

Video Processing in MATLAB . . . . . . . . . . . . . . . . . . . . . . . . .

2-3

Computer Vision System Toolbox Preferences . . . . . . . . . . .

Parallel Computing Toolbox Support . . . . . . . . . . . . . . . . . . .

2-4
2-4

Coordinate Systems
Coordinate Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Pixel Indices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Spatial Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3-D Coordinate Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3-2
3-2
3-2
3-3

vi

Contents

System Objects
What Is a System Toolbox? . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-2

What Are System Objects? . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-3

System Objects vs. MATLAB Functions . . . . . . . . . . . . . . . . .

System Objects vs. MATLAB Functions . . . . . . . . . . . . . . . .
Process Audio Data Using Only MATLAB Functions Code . .
Process Audio Data Using System Objects . . . . . . . . . . . . . .

4-5
4-5
4-5
4-6

System Design and Simulation in MATLAB . . . . . . . . . . . . . .

4-8

System Design and Simulation in Simulink . . . . . . . . . . . . . .

4-9

System Objects in MATLAB Code Generation . . . . . . . . . . .

System Objects in Generated Code . . . . . . . . . . . . . . . . . . .
System Objects in codegen . . . . . . . . . . . . . . . . . . . . . . . . .
System Objects in the MATLAB Function Block . . . . . . . . .
System Objects in the MATLAB System Block . . . . . . . . . .
System Objects and MATLAB Compiler Software . . . . . . . .

4-10
4-10
4-14
4-15
4-15
4-15

System Objects Methods That Support Code Generation . .

Code Generation Supported System Objects Methods . . . . .
Simulation-Only System Objects Methods . . . . . . . . . . . . . .

4-16
4-16
4-16

System Objects in Simulink . . . . . . . . . . . . . . . . . . . . . . . . . .

System Objects in the MATLAB Function Block . . . . . . . . .
System Objects in the MATLAB System Block . . . . . . . . . .

4-18
4-18
4-18

System Object Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

What Are System Object Methods? . . . . . . . . . . . . . . . . . . .
The Step Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Common Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-19
4-19
4-19
4-20

System Design in MATLAB Using System Objects . . . . . . . .

Predefined Components . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Create Components for Your System . . . . . . . . . . . . . . . . . .
Component Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Configure Component Property Values . . . . . . . . . . . . . . . .
Create and Configure Components at the Same Time . . . . .

4-22
4-22
4-22
4-23
4-23
4-24

Connecting System Objects . . . . . . . . . . . . . . . . . . . . . . . . .

Connect Components in a System . . . . . . . . . . . . . . . . . . . .
Run Your System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Reconfiguring Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Change a Tunable Property in Your System . . . . . . . . . . . .

4-25
4-25
4-27
4-28
4-28

System Design in Simulink Using System Objects . . . . . . .

Define New Kinds of System Objects for Use in Simulink . .
Test New System Objects in MATLAB . . . . . . . . . . . . . . . .
Add System Objects to Your Simulink Model . . . . . . . . . . .

4-31
4-31
4-36
4-37

Strategies for Real-Time Video Processing in

Simulink
Optimizing Your Implementation . . . . . . . . . . . . . . . . . . . . . .

5-2

Developing Your Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5-4

Data Type Support

Block Data Type Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6-2

Fixed-Point Support for MATLAB System Objects . . . . . . . .

Getting Information About Fixed-Point System Objects . . . . .
Setting System Object Fixed-Point Properties . . . . . . . . . . . .

6-3
6-3
6-4

vii

1
Product Overview

Product Overview

Computer Vision System Toolbox Product Description

Design and simulate computer vision and video processing systems
Computer Vision System Toolbox provides algorithms, functions, and apps for designing
and simulating computer vision and video processing systems. You can perform feature
detection, extraction, and matching; object detection and tracking; motion estimation;
and video processing. For 3-D computer vision, the system toolbox supports camera
calibration, stereo vision, 3-D reconstruction, and 3-D point cloud processing. With
machine learning based frameworks, you can train object detection, object recognition,
and image retrieval systems. Algorithms are available as MATLAB functions, System
objects, and Simulink blocks.
For rapid prototyping and embedded system design, the system toolbox supports fixedpoint arithmetic and C-code generation.

Key Features
Object detection and tracking, including the Viola-Jones, Kanade-Lucas-Tomasi
(KLT), and Kalman filtering methods
Training of object detection, object recognition, and image retrieval systems, including
cascade object detection and bag-of-features methods
Camera calibration for single and stereo cameras, including automatic checkerboard
detection and an app for workflow automation
Stereo vision, including rectification, disparity calculation, and 3-D reconstruction
3-D point cloud processing, including I/O, visualization, registration, denoising, and
geometric shape fitting
Feature detection, extraction, and matching
Support for C-code generation and fixed-point arithmetic with code generation
products

1-2

2
Computer Vision Algorithms and
Video Processing
Computer Vision Capabilities on page 2-2
Video Processing in MATLAB on page 2-3
Computer Vision System Toolbox Preferences on page 2-4

Computer Vision Algorithms and Video Processing

Computer Vision Capabilities

Computer Vision System Toolbox provides algorithms and tools for the design and
simulation of computer vision and video processing systems. The toolbox includes
algorithms for feature extraction, motion detection, object detection, object tracking,
stereo vision, video processing, and video analysis. Tools include video file I/O, video
display, drawing graphics, and compositing. Capabilities are provided as MATLAB
functions, MATLAB System objects, and Simulink blocks. For rapid prototyping and
embedded system design, the system toolbox supports fixed-point arithmetic and C-code
generation.
The link below provides an overview video of Computer Vision System Toolbox
capabilities and applications:
Computer Vision System Toolbox capabilities

2-2

Video Processing in MATLAB

Video Processing in MATLAB

Computer Vision System Toolbox provides algorithms and tools for video processing
workflows. You can read and write from common video formats, perform common video
processing algorithms such as deinterlacing and chroma-resampling, and display results
with text and graphics burnt in to the video. Video processing in MATLAB uses System
objects, which avoids excessive memory use by streaming data to and from video files.
The link below provides an introduction video to a typical workflow for motion
estimation:
Video processing in MATLAB

2-3

Computer Vision Algorithms and Video Processing

Computer Vision System Toolbox Preferences

To open Computer Vision System Toolbox preferences, on the Home tab, in the
Environment section, click
Toolbox.

Preferences. Select Computer Vision System

Parallel Computing Toolbox Support

Several Computer Vision System Toolbox functions support parallel computing using
multiple MATLAB workers. Select the Use Parallel check box to enable parallel
computing when possible.
2-4

Computer Vision System Toolbox Preferences

Parallel computing functionality requires a Parallel Computing Toolbox license and an

open MATLAB pool.
The functions and methods listed below take an optional logical input parameter,
'UseParallel' to control whether the individual function can use parfor. Set this
logical to 'true' to enable parallel processing for the function or method.
bagOfFeatures
encode (bagOfFeatures)
trainImageCategoryClassifier
imageCategoryClassifier
predict (imageCategoryClassifier)
See parpool for details on how to create a special job on a pool of workers, and connect
the MATLAB client to the parallel pool.

2-5

3
Coordinate Systems

Coordinate Systems

Coordinate Systems
You can specify locations in images using various coordinate systems. Coordinate
systems are used to place elements in relation to each other. Coordinates in pixel and
spatial coordinate systems relate to locations in an image. Coordinates in 3-D coordinate
systems describe the 3-D positioning and origin of the system.

Pixel Indices
Pixel coordinates enable you to specify locations in images. In the pixel coordinate
system, the image is treated as a grid of discrete elements, ordered from top to bottom
and left to right.

For pixel coordinates, the number of rows, r, downward, while the number of columns, c,
increase to the right. Pixel coordinates are integer values and range from 1 to the length
of the row or column. The pixel coordinates used in Computer Vision System Toolbox
software are one-based, consistent with the pixel coordinates used by Image Processing
Toolbox and MATLAB. For more information on the pixel coordinate system, see Pixel
Indices in the Image Processing Toolbox documentation.

Spatial Coordinates
Spatial coordinates enable you to specify a location in an image with greater granularity
than pixel coordinates. Such as, in the pixel coordinate system, a pixel is treated as a
discrete unit, uniquely identified by an integer row and column pair, such as (3,4). In
the spatial coordinate system, locations in an image are represented in terms of partial
pixels, such as (3.3, 4.7).
3-2

Coordinate Systems

For more information on the spatial coordinate system, see Spatial Coordinates in the
Image Processing Toolbox documentation.

3-D Coordinate Systems

When you reconstruct a 3-D scene, you can define the resulting 3-D points in one of two
coordinate systems. In a camera-based coordinate system, the points are defined relative
to the center of the camera. In a calibration pattern-based coordinate system, the points
are defined relative to a point in the scene.
The Computer Vision System Toolbox functions use the right-handed world coordinate
system. In this system, the x-axis points to the right, the y-axis points down, and the zaxis points away from the camera. To display 3-D points, use pcshow.
Camera-Based Coordinate System
Points represented in a camera-based coordinate system are described with the origin
located at the optical center of the camera.

Z
X

3-3

Coordinate Systems

In a stereo system, the origin is located at the optical center of Camera 1.

X
camera 1

camera 2

Y
When you reconstruct a 3-D scene using a calibrated stereo camera, the
reconstructScene and triangulate functions return 3-D points with the origin at
the optical center of Camera 1. When you use Kinect images, the depthToPointCloud
function returns 3-D points with the origin at the center of the RGB camera.
Calibration Pattern-Based Coordinate System
Points represented in a calibration pattern-based coordinate system are described with
the origin located at the (0,0) location of the calibration pattern.

X
Y

When you reconstruct a 3-D scene from multiple views containing a calibration pattern,
the resulting 3-D points are defined in the pattern-based coordinate system. The
Structure From Motion From Two Views example shows how to reconstruct a 3-D scene
from a pair of 2-D images containing a checkerboard pattern.

Related Examples

3-4

Measuring Planar Objects with a Calibrated Camera

Coordinate Systems

Structure From Motion From Two Views

Structure From Motion From Multiple Views

Depth Estimation From Stereo Video

3-5

4
System Objects
What Is a System Toolbox? on page 4-2
What Are System Objects? on page 4-3
System Objects vs. MATLAB Functions on page 4-5
System Design and Simulation in MATLAB on page 4-8
System Design and Simulation in Simulink on page 4-9
System Objects in MATLAB Code Generation on page 4-10
System Objects Methods That Support Code Generation on page 4-16
System Objects in Simulink on page 4-18
System Object Methods on page 4-19
System Design in MATLAB Using System Objects on page 4-22
System Design in Simulink Using System Objects on page 4-31

System Objects

What Is a System Toolbox?

System Toolbox products provide algorithms and tools for designing, simulating,
and deploying dynamic systems in MATLAB and Simulink. These toolboxes contain
MATLAB functions, System objects, and Simulink blocks that deliver the same design
and verification capabilities across MATLAB and Simulink, enabling more effective
collaboration among system designers. Available System Toolbox products include:
DSP System Toolbox
Communications System Toolbox
Computer Vision System Toolbox
Phased Array System Toolbox
System Toolboxes support floating-point and fixed-point streaming data simulation
for both sample- and frame-based data. They provide a programming environment for
defining and executing code for various aspects of a system, such as initialization and
reset. System Toolboxes also support code generation for a range of system development
tasks and workflows, such as:
Rapid development of reusable IP and test benches
Sharing of component libraries and systems models across teams
Large system simulation
C-code generation for embedded processors
Finite wordlength effects modeling and optimization
Ability to prototype and test on real-time hardware

4-2

What Are System Objects?

What Are System Objects?

A System object is a specialized kind of MATLAB object. System Toolboxes include
System objects and most System Toolboxes also have MATLAB functions and Simulink
blocks. System objects are designed specifically for implementing and simulating
dynamic systems with inputs that change over time. Many signal processing,
communications, and controls systems are dynamic. In a dynamic system, the values
of the output signals depend on both the instantaneous values of the input signals and
on the past behavior of the system. System objects use internal states to store that past
behavior, which is used in the next computational step. As a result, System objects are
optimized for iterative computations that process large streams of data, such as video
and audio processing systems.
For example, you could use System objects in a system that reads data from a file,
filters that data and then writes the filtered output to another file. Typically, a specified
amount of data is passed to the filter in each loop iteration. The file reader object uses
a state to keep track of where in the file to begin the next data read. Likewise, the file
writer object keeps tracks of where it last wrote data to the output file so that data is not
overwritten. The filter object maintains its own internal states to assure that the filtering
is performed correctly. This diagram represents a single loop of the system.

Many System objects support:

Fixed-point arithmetic (requires a Fixed-Point Designer license)
C code generation (requires a MATLAB Coder or Simulink Coder license)
HDL code generation (requires an HDL Coder license)
Executable files or shared libraries generation (requires a MATLAB Compiler
license)
Note: Check your product documentation to confirm fixed-point, code generation, and
MATLAB Compiler support for the specific System objects you want to use.
4-3

System Objects

In addition to the System objects provided with System Toolboxes, you can also create
your own System objects. See Define System Objects.

4-4

System Objects vs. MATLAB Functions

System Objects vs. MATLAB Functions

In this section...
System Objects vs. MATLAB Functions on page 4-5
Process Audio Data Using Only MATLAB Functions Code on page 4-5
Process Audio Data Using System Objects on page 4-6

System Objects vs. MATLAB Functions

Many System objects have MATLAB function counterparts. For simple, one-time
computations use MATLAB functions. However, if you need to design and simulate
a system with many components, use System objects. Using System objects is also
appropriate if your computations require managing internal states, have inputs that
change over time or process large streams of data.
Building a dynamic system with different execution phases and internal states using
only MATLAB functions would require complex programming. You would need code to
initialize the system, validate data, manage internal states, and reset and terminate
the system. System objects perform many of these managerial operations automatically
during execution. By combining System objects in a program with other MATLAB
functions, you can streamline your code and improve efficiency.

Process Audio Data Using Only MATLAB Functions Code

This example shows how to write MATLAB function-only code for reading audio data.
The code reads audio data from a file, filters it, and then plays the filtered audio data.
The audio data is read in frames. This code produces the same result as the System
objects code in the next example, allowing you to compare approaches.
Locate source audio file.
fname = 'speech_dft_8kHz.wav';

Obtain the total number of samples and the sampling rate from the source file.
audioInfo = audioinfo(fname);
maxSamples = audioInfo.TotalSamples;
fs = audioInfo.SampleRate;

4-5

System Objects

Define the filter to use.

b = fir1(160,.15);

Initialize the filter states.

z = zeros(1,numel(b)-1);

Define the amount of audio data to process at one time, and initialize the while loop
index.
frameSize = 1024;
nIdx = 1;

Define the while loop to process the audio data in processloop.

type processloop

while nIdx <= maxSamples(1)-frameSize+1

audio = audioread(fname,[nIdx nIdx+frameSize-1]);
[y,z] = filter(b,1,audio,z);
sound(y,fs);
nIdx = nIdx+frameSize;
end

The loop uses explicit indexing and state management, which can be a tedious and errorprone approach. You must have detailed knowledge of the states, such as, sizes and
data types. Another issue with this MATLAB-only code is that the sound function is not
designed to run in real time. The resulting audio is very choppy and barely audible.

Process Audio Data Using System Objects

This example shows how to write System objects code for reading audio data.
The code uses System objects from the DSP System Toolbox software to read audio
data from a file, filter it, and then play the filtered audio data. This code produces
the same result as the MATLAB code shown previously, allowing you to compare
approaches.
Locate source audio file.
fname = 'speech_dft_8kHz.wav';

4-6

System Objects vs. MATLAB Functions

Define the System object to read the file.

audioIn = dsp.AudioFileReader(fname,'OutputDataType','single');

Define the System object to filter the data.

filtLP = dsp.FIRFilter('Numerator',fir1(160,.15));

Define the System object to play the filtered audio data.

audioOut = audioDeviceWriter('SampleRate',audioIn.SampleRate);

Define the while loop to process the audio data.

while ~isDone(audioIn)
audio = step(audioIn);
y = step(filtLP,audio);
step(audioOut,y);
end

% Read audio source file

% Filter the data
% Play the filtered data

This System objects code avoids the issues present in the MATLAB-only code. Without
requiring explicit indexing, the file reader object manages the data frame sizes while the
filter manages the states. The audio device writer object plays each audio frame as it is
processed.

4-7

System Objects

System Design and Simulation in MATLAB

System objects allow you to design and simulate your system in MATLAB. You use
System objects in MATLAB as shown in this diagram.

4-8

Create individual components Create the System objects to use in your system.
See Create Components for Your System on page 4-22. In addition to the
System objects provided with System Toolboxes, you can also create your own
System objects. See Define System Objects.

Configure components If necessary, change the objects property values to model

your particular system. All System object properties have default values that you
may be able to use without changing them. See Component Properties on page
4-23.

Assemble components into system Write a MATLAB program that includes those
System objects, connecting them using MATLAB variables as inputs and outputs to
simulate your system. See Connecting System Objects on page 4-25.

Run the system Run your program, which uses the step method to run your
systems System objects. You can change tunable properties while your system is
running. See Run Your System on page 4-27 and Reconfiguring Objects on
page 4-28.

System Design and Simulation in Simulink

System Design and Simulation in Simulink

You can use System objects in your model to simulate in Simulink.
1

Create a System object to be used in your model. See Define New Kinds of System
Objects for Use in Simulink on page 4-31 for information.

Test your new System object in MATLAB. See Test New System Objects in
MATLAB on page 4-36

Add the System object to your model using the MATLAB System block. See Add
System Objects to Your Simulink Model on page 4-37 for information.

Add other Simulink blocks as needed and connect the blocks to construct your
system.

Run the system

4-9

System Objects

System Objects in MATLAB Code Generation

In this section...
System Objects in Generated Code on page 4-10
System Objects in codegen on page 4-14
System Objects in the MATLAB Function Block on page 4-15
System Objects in the MATLAB System Block on page 4-15
System Objects and MATLAB Compiler Software on page 4-15

System Objects in Generated Code

You can generate C/C++ code in MATLAB from your system that contains System objects
by using MATLAB Coder. You can generate efficient and compact code for deployment in
desktop and embedded systems and accelerate fixed-point algorithms.
System Objects Code with Persistent Objects for Code Generation
function ex_system_codegen
% Find corresponding interest points between a pair of images using local
% neighborhoods.
% Declare System objects as persistent.
persistent colorSpaceConverter
% Initialize persistent System objects only once
% Do this with 'if isempty(persistent variable).'
% This condition will be false after the first time.
if isempty(colorSpaceConverter)
% Create system objects. Pass property value arguments as constructor
% arguments. Property values must be constants during compile time.
colorSpaceConverter = vision.ColorSpaceConverter('Conversion',...
'RGB to intensity');
end
% Declare functions called into MATLAB that do not generate
% code as extrinsic.
coder.extrinsic('imread');
% The output of an extrinsic function is an mxArray - also called a MATLAB

4-10

System Objects in MATLAB Code Generation

% array. To use mxArrays returned by extrinsic functions, assign the

% mxArray to a variable whose type and size is defined.
imgLeft = zeros([300 400 3],'uint8');
imgRight = zeros([300 400 3],'uint8');
% Call extrinsic function
imgLeft = imread('viprectification_deskLeft.png');
imgRight = imread('viprectification_deskRight.png');
% Convert RGB to grayscale
I1 = step(colorSpaceConverter,imgLeft);
I2 = step(colorSpaceConverter,imgRight);
% Find corners
points1 = detectHarrisFeatures(I1);
points2 = detectHarrisFeatures(I2);
% Extract neighborhood features
[features1, valid_points1] = extractFeatures(I1,points1);
[features2, valid_points2] = extractFeatures(I2,points2);
% Match features
index_pairs = matchFeatures(features1, features2);
% Retrieve locations of corresponding points for each image
matchedPoints1 = valid_points1.Location(index_pairs(:,1),:);
matchedPoints2 = valid_points2.Location(index_pairs(:,2),:);
% Visualize corresponding points
coder.extrinsic('showMatchedFeatures')
figure;
showMatchedFeatures(I1,I2,matchedPoints1,matchedPoints2);

For another detailed code generation example, see Generate Code for MATLAB Handle
Classes and System Objects in the MATLAB Coder product documentation.
Usage Rules and Limitations for System Objects for Generating Code
The following usage rules and limitations apply to using System objects in code
generated from MATLAB.
Object Construction and Initialization
If objects are stored in persistent variables, initialize System objects once by
embedding the object handles in an if statement with a call to isempty( ).
4-11

System Objects

Set arguments to System object constructors as compile-time constants.

You cannot initialize System objects properties with other MATLAB class objects
as default values in code generation. You must initialize these properties in the
constructor.
Inputs and Outputs
System objects accept a maximum of 32 inputs. A maximum of 8 dimensions per input
is supported.
The data type of the inputs should not change.
If you want the size of inputs to change, verify that variable-size is enabled. Code
generation support for variable-size data also requires that the Enable variable
sizing option is enabled, which is the default in MATLAB.
Note: Variable-size properties in MATLAB Function block in Simulink are not
supported. System objects predefined in the software do not support variable-size if
their data exceeds the DynamicMemoryAllocationThreshold value.
Do not set System objects to become outputs from the MATLAB Function block.
Do not use the Save and Restore Simulation State as SimState option for any System
object in a MATLAB Function block.
Do not pass a System object as an example input argument to a function being
compiled with codegen.
Do not pass a System object to functions declared as extrinsic (functions called in
interpreted mode) using the coder.extrinsic function. System objects returned
from extrinsic functions and scope System objects that automatically become extrinsic
can be used as inputs to another extrinsic function, but do not generate code.
Tunable and Nontunable Properties
The value assigned to a nontunable property must be a constant and there can be at
most one assignment to that property (including the assignment in the constructor).
For most System objects, the only time you can set their nontunable properties during
code generation is when you construct the objects.
For System objects that are predefined in the software, you can set their tunable
properties at construction time or using dot notation after the object is locked.
For System objects that you define, you can change their tunable properties
at construction time or using dot notation during code generation. For
4-12

System Objects in MATLAB Code Generation

getNumInputsImpl and getNumOutputsImpl methods, if you set the

return argument from an object property, that object property must have the
Nontunable attribute.
Objects cannot be used as default values for properties.
Global Variables
Global variables are allowed in a System object, unless you will be using that System
object in Simulink via the MATLAB System block. To avoid syncing global variables
between a MEX file and the workspace, use a coder configuration object. For example:
f = coder.MEXConfig;
f.GlobalSyncMethod = 'NoSync'

Then, include '-config f' in your codegen command.

Methods
Code generation support is available only for these System object methods:
get
getNumInputs
getNumOutputs
isDone (for sources only)
isLocked
release
reset
set (for tunable properties)
step
For System objects that you define,
Code generation support is available only for these methods:
getDiscreteStateImpl
getNumInputsImpl
getNumOutputsImpl
infoImpl
isDoneImpl
4-13

System Objects

isInputDirectFeedThroughImpl
outputImpl
processTunedPropertiesImpl
releaseImpl Code is not generated automatically for the this method. To
release an object, you must explicitly call the release method in your code.
resetImpl
setupImpl
stepImpl
updateImpl
validateInputsImpl
validatePropertiesImpl
Code generation support for using dot notation depends on whether the System object
is predefined in the software or is one that you defined.
For System objects that are predefined in the software, you cannot use dot
notation to call methods.
For System objects that you define, you can use dot notation or function call
notation, with the System object as first argument, to call methods.

System Objects in codegen

You can include System objects in MATLAB code in the same way you include any
other elements. You can then compile a MEX file from your MATLAB code by using
the codegen command, which is available if you have a MATLAB Coder license. This
compilation process, which involves a number of optimizations, is useful for accelerating
simulations. See Getting Started with MATLAB Coder and MATLAB Classes for
more information.
Note: Most, but not all, System objects support code generation. Refer to the particular
objects reference page for information.

4-14

System Objects in MATLAB Code Generation

System Objects in the MATLAB Function Block

Using the MATLAB Function block, you can include any System object and any
MATLAB language function in a Simulink model. This model can then generate
embeddable code. System objects provide higher-level algorithms for code generation
than do most associated blocks. For more information, see What Is a MATLAB Function
Block? in the Simulink documentation.

System Objects in the MATLAB System Block

Using the MATLAB System block, you can include in a Simulink model individual
System objects that you create with a class definition file. The model can then generate
embeddable code. For more information, see MATLAB System Block in the Simulink
documentation.

System Objects and MATLAB Compiler Software

MATLAB Compiler software supports System objects for use inside MATLAB functions.
The compiler product does not support System objects for use in MATLAB scripts.

4-15

System Objects

System Objects Methods That Support Code Generation

In this section...
Code Generation Supported System Objects Methods on page 4-16
Simulation-Only System Objects Methods on page 4-16

Code Generation Supported System Objects Methods

Only the following methods are supported in code generation.
getDiscreteStateImpl
getNumInputsImpl
getNumOutputsImpl
isDoneImpl
infoImpl
isInputDirectFeedthroughImpl
outputImpl
processTunedPropertiesImpl
releaseImpl Code is not generated automatically for the this method. To release an
object, you must explicitly call the release method in your code.
resetImpl
setupImpl
stepImpl
updateImpl
validateInputsImpl
validatePropertiesImpl

Simulation-Only System Objects Methods

The following methods are for simulation only and do not support code generation.
getDiscreteStateSpecificationImpl
getHeaderImpl
4-16

System Objects Methods That Support Code Generation

getInputNamesImpl
getIconImpl
getOutputDataTypeImpl
getOutputNamesImpl
getOutputSizeImpl
getPropertyGroupsImpl
isInactivePropertyImpl
isOutputComplexImpl
loadObjectImpl
propagatedInputComplexity
propagatedInputDataType
propagatedInputFixedSize
propagatedInputSize
saveObjectImpl
supportsMultipleInstanceImpl

4-17

System Objects

System Objects in Simulink

In this section...
System Objects in the MATLAB Function Block on page 4-18
System Objects in the MATLAB System Block on page 4-18

System Objects in the MATLAB Function Block

You can include System object code in Simulink models using the MATLAB Function
block. Your function can include one or more System objects. Portions of your system may
be easier to implement in the MATLAB environment than directly in Simulink. Many
System objects have Simulink block counterparts with equivalent functionality. Before
writing MATLAB code to include in a Simulink model, check for existing blocks that
perform the desired operation.

System Objects in the MATLAB System Block

You can include individual System objects that you create with a class definition file
into Simulink using the MATLAB System block. This provides one way to add your
own algorithm blocks into your Simulink models. For information, see System Object
Integration in the Simulink documentation.

4-18

System Object Methods

System Object Methods

In this section...
What Are System Object Methods? on page 4-19
The Step Method on page 4-19
Common Methods on page 4-20

What Are System Object Methods?

After you create a System object, you use various object methods to process data or
obtain information from or about the object. All methods that are applicable to an object
are described in the reference pages for that object. System object method names begin
with a lowercase letter and class and property names begin with an uppercase letter.
The syntax for using methods is <method>(<handle>), such as step(H), plus possible
extra input arguments.
System objects use a minimum of two commands to process dataa constructor to
create the object and the step method to run data through the object. This separation
of declaration from execution lets you create multiple, persistent, reusable objects,
each with different settings. Using this approach avoids repeated input validation
and verification, allows for easy use within a programming loop, and improves overall
performance. In contrast, MATLAB functions must validate parameters every time you
call the function.
These advantages make System objects particularly well suited for processing streaming
data, where segments of a continuous data stream are processed iteratively. This ability
to process streaming data provides the advantage of not having to hold large amounts of
data in memory. Use of streaming data also allows you to use simplified programs that
use loops efficiently.

The Step Method

The step method is the key System object method. You use step to process data using
the algorithm defined by that object. The step method performs other important tasks
related to data processing, such as initialization and handling object states. Every
System object has its own customized step method, which is described in detail on the
step reference page for that object. For more information about the step method and
other available methods, see the descriptions in Common Methods on page 4-20.
4-19

System Objects

Common Methods
All System objects support the following methods, each of which is described in a method
reference page associated with the particular object. In cases where a method is not
applicable to a particular object, calling that method has no effect on the object.
Method

Description

step

Processes data using the algorithm defined by the object. As

part of this processing, it initializes needed resources, returns
outputs, and updates the object states. After you call the
step method, you cannot change any input specifications (i.e.,
dimensions, data type, complexity). During execution, you can
change only tunable properties. The step method returns
regular MATLAB variables.
Example: Y = step(H,X)

4-20

release

Releases any special resources allocated by the object, such

as file handles and device drivers, and unlocks the object.
For System objects, use the release method instead of a
destructor.

reset

Resets the internal states of a locked object to the initial values

for that object and leaves the object locked

getNumInputs

Returns the number of inputs (excluding the object itself)

expected by the step method. This number varies for an object
depending on whether any properties enable additional inputs.

getNumOutputs

Returns the number of outputs expected from the step

method. This number varies for an object depending on
whether any properties enable additional outputs.

getDiscreteState

Returns the discrete states of the object in a structure. If the

object is unlocked (when the object is first created and before
you have run the step method on it or after you have released
the object), the states are empty. If the object has no discrete
states, getDiscreteState returns an empty structure.

clone

Creates another object of the same type with the same property
values

isLocked

Returns a logical value indicating whether the object is locked.

System Object Methods

Method

Description

isDone

Applies to source objects only. Returns a logical value

indicating whether the step method has reached the end of
the data file. If a particular object does not have end-of-data
capability, this method value returns false.

info

Returns a structure containing characteristic information

about the object. The fields of this structure vary depending on
the object. If a particular object does not have characteristic
information, the structure is empty.

For a complete list of methods for writing new System objects, see System Objects
Methods for Defining New Objects.

4-21

System Objects

System Design in MATLAB Using System Objects

In this section...
Predefined Components on page 4-22
Create Components for Your System on page 4-22
Component Properties on page 4-23
Configure Component Property Values on page 4-23
Create and Configure Components at the Same Time on page 4-24
Connecting System Objects on page 4-25
Connect Components in a System on page 4-25
Run Your System on page 4-27
Reconfiguring Objects on page 4-28
Change a Tunable Property in Your System on page 4-28

Predefined Components
The example in the next section shows how to use System objects that are predefined in
the software. If you use a function to create and use a System object, specify the object
creation using conditional code. This will prevent errors if that function is called within a
loop.
You can also create your own System objects (see Define System Objects).

Create Components for Your System

This section of the example shows how to set up your system. The predefined components
you need are:
vision.VideoFileReader Read the file of video data
vision.AlphaBlender Overlay edges onto the original video images
vision.VideoPlayer Play the video
The example uses a function for edge detection. First, create the component objects, using
default property settings. Create three VideoPlayer objects. One to play the original
video, one to play the edges, and one to show the edges overlaid on the original video.
4-22

System Design in MATLAB Using System Objects

hVideoSrc = vision.VideoFileReader;
hAB = vision.AlphaBlender;
hVideoOrig = vision.VideoPlayer;
hVideoEdges = vision.VideoPlayer;
hVideoOverlay = vision.VideoPlayer;

Component Properties
When to Configure Components
If you did not set an object's properties when you created it and do not want to use
default values, you must explicitly set those properties. Some properties allow you to
change their values while your system is running. See Reconfiguring Objects on page
4-28 for information.
Most properties are independent of each other. However, some System object properties
enable or disable another property or limit the values of another property. To avoid
errors or warnings, you should set the controlling property before setting the dependent
property.
Display Component Property Values
To display the current property values for an object, type that objects handle name at
the command line (such as audioIn). To display the value of a specific property, type
objecthandle.propertyname (such as audioIn.FileName).

Configure Component Property Values

This section of the example shows how to configure the components for your system by
setting the component objects properties.
Use this procedure if you have created your components separately from configuring
them. You can also create and configure your components at the same time, as described
in a later example.
For the video file reader object, specify the file to read and set the image color space.
For the alpha blender object, specify the type of operation to use.
For the video player objects, specify the names, the window size, and the window
position.
4-23

System Objects

hVideoSrc.Filename = 'vipmen.avi';
hVideoSrc.ImageColorSpace = 'Intensity';
hAB.Operation = 'Highlight selected pixels';
WindowSize = [190 150];
hVideoOrig.Name = 'Original';
hVideoOrig.Position = [10 hVideoOrig.Position(2) WindowSize];
hVideoEdges.Name = 'Edges';
hVideoEdges.Position = [210 hVideoOrig.Position(2) WindowSize];
hVideoOverlay.Name = 'Overlay';
hVideoOverlay.Position = [410 hVideoOrig.Position(2) WindowSize];

Create and Configure Components at the Same Time

This section of the example shows how to create your System object components and
configure the desired properties at the same time. Specify each property with a 'Name',|
Value| argument pair.
For the video file reader object, specify the file to read and set the image color space.
For the alpha blender object, specify the type of operation to use.
For the video player objects, specify the names, the window size, and the window
position.
hVideoSrc = vision.VideoFileReader('vipmen.avi', ...
'ImageColorSpace', 'Intensity');
hAB = vision.AlphaBlender('Operation', 'Highlight selected pixels');
WindowSize = [190 150];
hVideoOrig = vision.VideoPlayer('Name', 'Original',...
'Position',[10 100 WindowSize]);
hVideoEdges = vision.VideoPlayer('Name', 'Edges',...
'Position',[210 100 WindowSize]);
hVideoOverlay = vision.VideoPlayer('Name', 'Overlay',...
'Position',[410 100 WindowSize]);

4-24

System Design in MATLAB Using System Objects

Connecting System Objects

After you have determined the components you need and have created and configured
your System objects, assemble your system. You use the System objects like other
MATLAB variables and include them in MATLAB code. You can pass MATLAB variables
into and out of System objects.
The main difference between using System objects and using functions is the step
method. The step method is the processing command for each System object and is
customized for that specific System object. This method initializes your objects and
controls data flow and state management of your system. You typically use step within a
loop.
You use the output from an objects step method as the input to another objects step
method. For some System objects, you can use properties of those objects to change the
number of inputs or outputs. To verify that the appropriate number of input and outputs
are being used, you can use getNumInputs and getNumOutputs on any System object.
For information on all available System object methods, see System Object Methods on
page 4-19.

Connect Components in a System

This section shows how to connect the components together to read the file, detect edges,
form a composite image, and display the video data. The while loop uses the isDone
method to read through the entire file. The step method is used on each object inside the
loop.
while ~isDone(hVideoSrc)
frame
= step(hVideoSrc);
edges
= edge(frame,'Prewitt',15/256);
composite = step(hAB, frame, edges);
step(hVideoOrig,frame);
step(hVideoEdges,edges);
step(hVideoOverlay,composite);

% Read input video

% Edge detection
% AlphaBlender
% Display original
% Display edges
% Display edges overlaid

end
release(hVideoSrc);

4-25

System Objects

4-26

System Design in MATLAB Using System Objects

Run Your System

How to Run Your System
Run your code either by typing directly at the command line or running a file containing
your program. When you run the code for your system, the step method instructs each
object to process data through that object.
What You Cannot Change While Your System Is Running
The first call to the step method initializes and then locks your object. When a System
object has started processing data, it is locked to prevent changes that would disrupt its
processing. Use the isLocked method to verify whether an object is locked. When the
object is locked, you cannot change:
Number of inputs or outputs
Data type of inputs or outputs
Data type of any tunable property
Dimensions of inputs or tunable properties, except for System objects that support
variable-size data
Value of any nontunable property
To make changes to your system while it is running, see Reconfiguring Objects on page
4-28.
4-27

System Objects

Reconfiguring Objects
When Can You Change Component Properties?
When a System object has started processing data, it is locked to prevent changes that
would disrupt its processing. You can use isLocked on any System object to verify
whether it is locked or not. When processing is complete, you can use the release
method to unlock a System object.
Some object properties are tunable, which enables you to change them even if the object
is locked. Unless otherwise specified, System objects properties are nontunable. Refer
to the objects reference page to determine whether an individual property is tunable.
Typically, tunable properties are not critical to how the System object processes data.
Change Input Complexity or Dimensions
During simulation, some System objects do not allow complex data if the object was
initialized with real data. You cannot change any input complexity during code
generation.
You can change the value of a tunable property without a warning or error being
produced. For all other changes at run time, an error occurs.

Change a Tunable Property in Your System

You can change the threshold value as your code is running by modifying an object
property. This example modifies the Location property of the AlphaBlender object.
The change takes effect the next time the step method is called (in this case, at the next
iteration of the while loop).
while ~isDone(hVideoSrc)
frame
= step(hVideoSrc);
edges
= edge(frame,'Prewitt',15/256);
composite = step(hAB, frame, edges);

4-28

% Read input video

% Edge detection
% AlphaBlender

hAB.Location(1) = hAB.Location(1)+1;
if (hAB.Location(1)==120)
hAB.Location = [1 1];
end

% Offset edges

step(hVideoOrig,frame);

% Display original

% wrap

System Design in MATLAB Using System Objects

step(hVideoEdges,edges);
step(hVideoOverlay,composite);

% Display edges
% Display edges overlaid

end

4-29

System Objects

4-30

System Design in Simulink Using System Objects

System Design in Simulink Using System Objects

In this section...
Define New Kinds of System Objects for Use in Simulink on page 4-31
Test New System Objects in MATLAB on page 4-36
Add System Objects to Your Simulink Model on page 4-37

Define New Kinds of System Objects for Use in Simulink

Define System Object with Block Customizations on page 4-31
Define System Object with Nondirect Feedthrough on page 4-34
A System object is a component you can use to create your system in MATLAB. You can
write the code in MATLAB and use that code to create a block in Simulink. To define
your own System object, you write a class definition file, which is a text-based MATLAB
file that contains the code defining your object. See System Object Integration in the
Simulink documentation.
Define System Object with Block Customizations
This example shows how to create a System object for use in Simulink. The example
performs system identification using a least mean squares (LMS) adaptive filter and is
similar to the System Identification Using MATLAB System Blocks Simulink example.
This example shows how to create a class definition text file to define your System
object. The code in this example creates a least mean squares (LMS) filter and includes
customizations to the block icon and dialog appearance.
Note: Instead of manually creating your class definition file, you can use the New >
System Object > Simulink Extension menu option to open a template. This template
includes customizations of the System object for use in the Simulink MATLAB System
block. You edit the template file, using it as guideline, to create your own System object.
On the first line of the class definition file, specify the name of your System object and
subclass from both matlab.System and matlab.system.mixin.CustomIcon. The
matlab.System base class enables you to use all the basic System object methods and
specify the block input and output names, title, and property groups. The CustomIcon
mixin class enables the method that lets you specify the block icon.
4-31

System Objects

Add the appropriate basic System object methods to set up, reset, set the number of
inputs and outputs, and run your algorithm. See the reference pages for each method and
the full class definition file below for the implementation of each of these methods.
Use the setupImpl method to perform one-time calculations and initialize variables.
Use the stepImpl method to implement the blocks algorithm.
Use the resetImpl method to reset the state properties or DiscreteState
properties.

Use the getNumInputsImpl and getNumOutputsImpl methods to specify the

number of inputs and outputs, respectively.
Add the appropriate CustomIcon methods to define the appearance of the MATLAB
System block in Simulink. See the reference pages for each method and the full class
definition file below for the implementation of each of these methods.
Use the getHeaderImpl method to specify the title and description to display on the
block dialog.
Use the getPropertyGroupsImpl method to specify groups of properties to display
on the block dialog.
Use the getIconImpl method to specify the text to display on the block icon.

Use the getInputNamesImpl and getOutputNamesImpl methods to specify the

labels to display for the block input and output ports.
The full class definition file for the least mean squares filter is:
classdef lmsSysObj < matlab.System &...
matlab.system.mixin.CustomIcon
% lmsSysObj Least mean squares (LMS) adaptive filtering.
% #codegen
properties
% Mu Step size
Mu = 0.005;
end
properties (Nontunable)
% Weights Filter weights
Weights = 0;
% N Number of filter weights
N = 32;
end

4-32

System Design in Simulink Using System Objects

properties (DiscreteState)
X;
H;
end
methods (Access = protected)
function setupImpl(obj)
obj.X = zeros(obj.N,1);
obj.H = zeros(obj.N,1);
end
function [y, e_norm] = stepImpl(obj,d,u)
tmp = obj.X(1:obj.N-1);
obj.X(2:obj.N,1) = tmp;
obj.X(1,1) = u;
y = obj.X'*obj.H;
e = d-y;
obj.H = obj.H + obj.Mu*e*obj.X;
e_norm = norm(obj.Weights'-obj.H);
end
function resetImpl(obj)
obj.X = zeros(obj.N,1);
obj.H = zeros(obj.N,1);
end
end
% Block icon and dialog customizations
methods (Static, Access = protected)
function header = getHeaderImpl
header = matlab.system.display.Header(...
'lmsSysObj', ...
'Title', 'LMS Adaptive Filter');
end
function groups = getPropertyGroupsImpl
upperGroup = matlab.system.display.SectionGroup(...
'Title','General',...
'PropertyList',{'Mu'});
lowerGroup = matlab.system.display.SectionGroup(...
'Title','Coefficients', ...
'PropertyList',{'Weights','N'});

4-33

System Objects

groups = [upperGroup,lowerGroup];
end
end
methods (Access = protected)
function icon = getIconImpl(~)
icon = sprintf('LMS Adaptive\nFilter');
end
function [in1name, in2name] = getInputNamesImpl(~)
in1name = 'Desired';
in2name = 'Actual';
end
function [out1name, out2name] = getOutputNamesImpl(~)
out1name = 'Output';
out2name = 'EstError';
end
end
end

Define System Object with Nondirect Feedthrough

This example shows how to create a System object for use in Simulink. The example
performs system identification using a least mean squares (LMS) adaptive filter and
uses feedback loops. It is similar to the System Identification Using MATLAB System
Blocks Simulink example. For information on feedback loops, see Use System Objects in
Feedback Loops.
This example shows how to create a class definition text file to define your System object.
The code in this example creates an integer delay and includes customizations to the
block icon. It implements a System object that you can use for nondirect feedthrough.
On the first line of the class definition file, subclass from matlab.System and
matlab.system.mixin.CustomIcon. The matlab.System base class enables you to
use all the basic System object methods and specify the block input and output names,
title, and property groups. The CustomIcon mixin class enables the method that lets you
specify the block icon. The Nondirect mixin enables the methods that let you specify
how the block is updated and what it outputs.
Add the appropriate basic System object methods to set up and reset the object and set
and validate the properties. Since this object supports nondirect feedthrough, you do not
implement the stepImpl method. You implement the updateImpl and outputImpl
4-34

System Design in Simulink Using System Objects

methods instead. See the reference pages for each method and the full class definition file
below for the implementation of each of these methods.
Use the setupImpl method to initialize some of the objects properties.
Use the resetImpl method to reset the property states.
Use the validatePropertiesImpl method to check that the property values are
valid.
Add the following Nondirect mixin class methods instead of the stepImpl method to
specify how the block updates its state and its output. See the reference pages and the
full class definition file below for the implementation of each of these methods.
Use the outputImpl method to implement code to calculate the block output.
Use the updateImpl method to implement code to update the blocks internal states.
Use the isInputDirectFeedthroughImpl method to specify that the block is not
direct feedthrough. Its inputs do not directly affect its outputs.
Add the getIconImpl method to define the block icon when it is used in Simulink via
the MATLAB System block. See the reference page and the full class definition file below
for the implementation of this method.
The full class definition file for the delay is:
classdef intDelaySysObj < matlab.System &...
matlab.system.mixin.Nondirect &...
matlab.system.mixin.CustomIcon
% intDelaySysObj Delay input by specified number of samples.
% #codegen
properties
% InitialOutput Initial output
InitialOutput = 0;
end
properties (Nontunable)
% NumDelays Number of delays
NumDelays = 1;
end
properties (DiscreteState)
PreviousInput;
end

4-35

System Objects

methods (Access = protected)

function setupImpl(obj, ~)
obj.PreviousInput = ones(1,obj.NumDelays)*obj.InitialOutput;
end
function [y] = outputImpl(obj, ~)
% Output does not directly depend on input
y = obj.PreviousInput(end);
end
function updateImpl(obj, u)
obj.PreviousInput = [u obj.PreviousInput(1:end-1)];
end
function flag = isInputDirectFeedthroughImpl(~,~)
flag = false;
end
function validatePropertiesImpl(obj)
if ((numel(obj.NumDelays)>1) || (obj.NumDelays <= 0))
error('Number of delays must be positive non-zero scalar value.');
end
if (numel(obj.InitialOutput)>1)
error('Initial output must be scalar value.');
end
end
function resetImpl(obj)
obj.PreviousInput = ones(1,obj.NumDelays)*obj.InitialOutput;
end
function icon = getIconImpl(~)
icon = sprintf('Integer\nDelay');
end
end
end

Test New System Objects in MATLAB

1

Create an instance of your new System object. For example, create an instance of the
lmsSysObj.
s = lmsSysObj;

4-36

System Design in Simulink Using System Objects

Run the step method on the object multiple times with different inputs. This tests for
syntax errors and other possible issues before you add it to Simulink. For example,
desired = 0;
actual = 0.2;
step(s,desired,actual);

Add System Objects to Your Simulink Model

1
2
3

Add your System objects to your Simulink model by using the MATLAB System block
as described in Mapping System Objects to Block Dialog Box.
Add other Simulink blocks, connect them, and configure any needed parameters to
complete your model as described in the Simulink documentation. See the System
Identification for an FIR System Using MATLAB System Blocks Simulink example.
Run your model in the same way you run any Simulink model.

4-37

5
Strategies for Real-Time Video
Processing in Simulink
Optimizing Your Implementation on page 5-2
Developing Your Models on page 5-4

Strategies for Real-Time Video Processing in Simulink

Optimizing Your Implementation

Video processing is computationally intensive, and the ability to perform real-time video
processing is affected by the following factors:
Hardware capability
Model complexity
Model implementation
Input data size
Optimizing your implementation is a crucial step toward real-time video processing. The
following tips can help improve the performance of your model:
Minimize the number of blocks in your model.
Process only the regions of interest to reduce the input data size.
Use efficient algorithms or the simplest version of an algorithm that achieves the
desired result.
Use efficient block parameter settings. However, you need to decide whether these
settings best suit your algorithm. For example, the most efficient block parameter
settings might not yield the most accurate results. You can find out more about
individual block parameters and their effect on performance by reviewing specific
block reference pages.
The two following examples show settings that make each block's operation the least
computationally expensive:
Resize block Interpolation method = Nearest neighbor
Blocks that support fixed point On the Fixed-Point tab, Overflow mode =
Wrap
Choose data types carefully.
Avoid data type conversions.
Use the smallest data type necessary to represent your data to reduce memory
usage and accelerate data processing.
In simulation mode, models with floating-point data types run faster than models
with fixed-point data types. To speed up fixed-point models, you must run them
in accelerator mode. Simulink contains additional code to process all fixed-point
5-2

Optimizing Your Implementation

data types. This code affects simulation performance. After you run your model
in accelerator mode or generate code for your target using the Simulink Coder,
the fixed-point data types are specific to the choices you made for the fixed-point
parameters. Therefore, the fixed-point model and generated code run faster.

5-3

Strategies for Real-Time Video Processing in Simulink

Developing Your Models

Use the following general process guidelines to develop real-time video processing models
to run on embedded targets. By optimizing the model at each step, you improve its final
performance.
1

Create the initial model and optimize the implementation algorithm. Use floatingpoint data types so that the model runs faster in simulation mode. If you are working
with a floating-point processor, go to step 3.

If you are working with a fixed-point processor, gradually change the model data
types to fixed point, and run the model after every modification.
During this process, you can use data type conversion blocks to isolate the
floating point sections of the model from the fixed-point sections. You should see a
performance improvement if you run the model in accelerator mode.

5-4

Remove unnecessary sink blocks, including scopes, and blocks that log data to files.

Compile the model for deployment on the embedded target.

6
Data Type Support

Data Type Support

Block Data Type Support

The Computer Vision System Toolbox Data Type Support Table is available through the
Simulink model Help menu. The table provides information about data type support and
code generation coverage for all Computer Vision System Toolbox blocks. Select Help >
Simulink> Block Data Types & Code Generation Support > Computer Vision
System Toolbox.

6-2

Fixed-Point Support for MATLAB System Objects

Fixed-Point Support for MATLAB System Objects

In this section...
Getting Information About Fixed-Point System Objects on page 6-3
Setting System Object Fixed-Point Properties on page 6-4
For information on working with Fixed-Point features, refer to the Fixed-Pointtopic.

Getting Information About Fixed-Point System Objects

System objects that support fixed-point data processing have fixed-point properties.
When you display the properties of a System object, click Show all properties
at the end of the property list to display the fixed-point properties for that object.
You can also display the fixed-point properties for a particular object by typing
vision.<ObjectName>.helpFixedPoint at the command line.
The following Computer Vision System Toolbox objects support fixed-point data
processing.
Fixed-Point Data Processing Support
vision.AlphaBlender
vision.Autocorrelator
vision.BlobAnalysis
vision.BlockMatcher
vision.Convolver
vision.Crosscorrelator
vision.DCT
vision.Deinterlacer
vision.DemosaicInterpolator
vision.FFT
vision.HoughLines
vision.IDCT
vision.IFFT
vision.MarkerInserter
vision.Maximum
vision.Mean
vision.Median
vision.Minimum
6-3

Data Type Support

vision.Pyramid
vision.ShapeInserter
vision.Variance

Setting System Object Fixed-Point Properties

Several properties affect the fixed-point data processing used by a System object. Objects
perform fixed-point processing and use the current fixed-point property settings when
they receive fixed-point input.
You change the values of fixed-point properties in the same way as you change any
System object property value. You also use the Fixed-Point Designer numerictype
object to specify the desired data type as fixed point, the signedness, and the word- and
fraction-lengths.
In the same way as for blocks, the data type properties of many System objects can set
the appropriate word lengths and scalings automatically by using full precision. System
objects assume that the target specified on the Configuration Parameters Hardware
Implementation target is ASIC/FPGA.
If you have not set the property that activates a dependent property and you attempt
to change that dependent property, a warning message displays. For example, for
the vision.EdgeDetector object, before you set CustomProductDataType to
numerictype(1,16,15) set ProductDataType to 'Custom'.
Note: System objects do not support fixed-point word lengths greater than 128 bits.
For any System object provided in the Toolbox, the fimath settings for any fimath
attached to a fi input or a fi property are ignored. Outputs from a System object never
have an attached fimath.

6-4