WipFrag User's Manual PDF

WipFrag 2.
6
With Optional WipJoint Module
Operating Manual
WipFrag 2.6 Operating Manual
Disclaimer
WipWare image analysis software is the end product of more than ten years of research
and has been developed and tested in a broad range of applications. Every effort is
made to supply a reliable product backed by prompt and efficient customer support.
WipWare Inc. makes no representations or warranties with respect to the contents of
this document or the performance or results of this software, and specifically disclaims
any implied warrantees, of any kind, oral or written. In the event any warranties are
deemed to apply, the sole remedy for breech of such warranty and the sole obligation of
WipWare Inc. is the repair or replacement of the document and/or software, at the
discretion of WipWare Inc.
Copyright
WipFrag© and WipJoint© are the property of WipWare Inc.
WipWare Customer Support:
We welcome your feedback and suggestions for improving and extending WipWare.
Questions concerning WipWare sales, routine operation, user training, and blasting
applications should be addressed to:
Tom Palangio,
145 Palangio Rd.,
Bonfield Ontario,
P0H-1E0.
Fax: (705) 776-1580
Phone: (705) 776-1882
WipWare Technical Support:
Questions concerning software performance should be addressed to:
Norbert Maerz,
145 Palangio Rd.,
Bonfield Ontario,
P0H-1E0.
Fax: (705) 776-1580
Phone: (705) 776-1882
2
TABLE OF CONTENTS
TABLE OF CONTENTS 3
1. INTRODUCTION 6
1.1 FRAGMENTATION MEASUREMENT 6
1.2 THE WIPFRAG METHOD 6
2. SETTING UP THE WIPFRAG COMPUTER 8
2.1 PREPARATIONS 8
2.1.1 Desktop Workstation Installation 8
2.1.3 Other Preparations 8
3. WIPFRAG OPERATIONS 9
3.1 WIPFRAG WINDOW 9
3.2 WIPFRAG OPERATIONS 10
3.3.1 Basic Operation 10
3.3.2 Basic Setup 10
4. WIPFRAG ACTION MENUS 11
4.1 MAIN MENUS AND TOOLBARS 11
4.1.1 Opening Menu 11
4.1.2 Image View Menu 11
4.1.3 Image View Tool Bar 11
4.2 FILE MENU AND TOOLS 12
4.2.1 Open 12
4.2.2 Open Multiple 12
4.2.3 Capture New Image 12
4.2.4 Close 13
4.2.5 Save 14
4.2.6 Save As 14
4.2.7 Open Net 14
4.2.8 Save Net 14
4.2.9 Print 14
4.2.10 Print Preview 14
4.2.11 Print Setup 14
4.2.12 (Recently used file list) 14
4.2.13 Exit 14
4.3 EDIT MENU AND TOOLS 15
4.3.1 Undo 15
4.3.2 Draw Line – F2 15
4.3.3 Draw Polyline – F3 15
3
4.3.4 Block Delete – F5 16

4.3.5 Delete Line – F4 16
4.3.6 Snap to 16
4.3.7 Ignore Block – F6 16
4.3.8 Mark as Fines – F7 16
4.4 VIEW MENU AND TOOLS 17
4.4.1 Edit Net 17
4.4.2 Edit Exclusion Zone 17
4.4.3 Zoom In 17
4.4.4 Zoom Out 17
4.4.5 Status Bar 17
4.5 FRAGMENTATION MENU 18
4.5.1 Set Scale 18
4.5.3 Generate Net 20
4.5.4 Sieve 21
4.6 WIPJOINT MENU 22
4.6.1 Set Scale 22
4.6.2 Generate Edges 22
4.6.3 Generate Polygons 22
4.6.4 Edit Edges 22
4.6.5 Edit Polygons 22
4.6.6 Analyze Joints 23
4.7 RIGHT CLICK ON IMAGE 24
4.7.1 Zoom In 24
4.7.2 Zoom Out 24
4.7.3 Ignore This Block – F6 24
4.7.4 Mark As Fines – F7 24
4.10. OPTIONS MENUS 25
4.10.1 Edge Detection Parameters 25
4.10.1.1 Edge Detection Settings 26
4.11 CANNY PARAMETERS 28
4.11.1 Canny Edge Detection Settings 29
4.12 SIEVE OPTIONS 30
4.12 WINDOW MENU 34
4.12.1 New Window 34
4.12.2 Cascade 34
4.12.3 Tile 34
4.12.4 Arrange Icons 34
4.13. HELP MENU 35
4.13.1 Help Topics 35
4.13.2 About WipFrag 35
APPENDIX 1: WIPFRAG “FINES” STRATEGIES 36
4
Engineering Solution 36
Empirically Calibrated Solution 36
Zoom-Merge Solution 36
APPENDIX 2: ROCKPILE SAMPLING AND PHOTOGRAPHY 37
Rock pile Sampling 37
Photography and Lighting 37
APPENDIX 3: CALIBRATION SAMPLES AND TECHNIQUES 39
APPENDIX 4: DEFINITIONS OF WIPFRAG STATISTICS 41
APPENDIX 5: FILE NAMING AND GRAPH LABELING CONVENTIONS 44
APPENDIX 6: WIPFRAG SECURITY KEY 45
APPENDIX 7: EXAMPLES OF WIPFRAG OUTPUT 46
APPENDIX 8: IMAGING MOVING PICTURES 47

Interlaced Video 47
APPENDIX 9: CAMERA ASSIGNMENTS 48
APPENDIX 10: OPTIONAL WIPJOINT MODULE 49

WIPJOINT 49
JOINTING MEASUREMENT 49
THE WIPJOINT METHOD 50
JOINT MEASUREMENT 50
JOINT GRAPHS 51
APPENDIX 11: OPTIONAL WIPJOINT IMAGE SAMPLING 52
Enhancement of Joints on the Face 52
Photography and Lighting 52
APPENDIX 12: OPTIONAL WIPJOINT OUTPUT 53
APPENDIX 13: WDM DRIVERS/V. 2.4 INSTALLATION 54

VERSION 2.4 INSTALLATION 54
RUNNING VERSION 2.4 FOR THE FIRST TIME. 54
APPENDIX 14: EXAMPLES OF WIPFRAG OUTPUTS 56
5
1. INTRODUCTION
1.1 Fragmentation Measurement
Blast fragmentation measurement is a recent addition to an extensive family of

granulometry applications. However, whereas small particles can be measured in the
laboratory by sieving, centrifugation, sedimentation, or optical microscopy, only digital
optical (photoanalysis) methods are of any practical value when the fragments are much
larger, as in blasting. The introduction of photoanalysis as a viable and practical tool has
had to await recent dramatic improvements in the speed, performance and cost of video
cameras and microcomputers, also, the development of software systems that can
recognize fragment edges in the muck pile, and discriminate one fragment from the
next.
1.2 The WipFrag Method
WipFrag starts with a video image of broken rock and converts it into a "net" of
fragments. Then it measures the net and displays and plots a selection of fragmentation
statistics and graphs. WipFrag allows you to compare the resulting net against the rock
image and to correct inaccuracies by manual editing with a mouse.
WipFrag is designed for two main types of input:
Roving camera input employs a hand-held camera. A robust and lightweight video
camcorder is recommended in most applications. Videotape provides compact storage
for any number of images, eliminates delays for processing of film, and allows you to
combine images at different magnifications for improved accuracy and resolution. The
WipFrag workstation also accepts input in the form of photo prints, 35 mm slides, digital
camera disks, or live images of crushed rock. Using images received by courier or mail,
a single central workstation can provide a service to any number of operators, not only
for routine processing of fragmentation data, but also for user training, calibration and
research.
Please refer to Appendix 2 for all image acquisition tips.
On-line monitoring uses fixed video cameras focused on draw points, conveyors, etc.
Images are automatically captured at preset intervals in scan mode or triggered by
motion sensors when a truck passes beneath the camera. They are transmitted to a
workstation in the mine office, and the results are displayed as strip-charts or graphs
with acceptance limits for size fractions including oversize and fines.
6
This manual describes use of a WipFrag workstation with roving camera input. On-line
monitoring requires customized hardware to protect cameras and lighting against dust
and blast damage and to avoid image blur when monitoring rock on a moving conveyor.
If you interested in online monitoring, please contact WipWare Inc. for more information
on the Momentum (Conveyor) and Reflex (Vehicle) monitoring systems.
7
2. SETTING UP THE WIPFRAG COMPUTER
2.1 Preparations
2.1.1 Desktop Workstation Installation

The desktop workstation requires any modern PC or notebook. This system utilizes a
DFG/LC4 framegrabber card, an Imperx PCMCIA card or firewire inputs for video.
2.1.3 Other Preparations

Assemble a copy stand alongside the computer. For initial SETUP and trials of
WipFrag, place a sample rock pile photo or crushed rock sample on a copy stand base
along with scale bar. Keep the print flat with the glass plate or masking tape, and adjust
the lamps to give uniform illumination (identical 30 to 60 W light bulbs in each of the
lamps) with no glare due to reflections.
Attach the camera/camcorder to the moving arm on the pillar of the copy stand and
connect it to the computer.
Power up the computer and install and start WipFrag.

For easy install procedures please follow the directions found in the WipFrag Quickstart
guide in the CD box.
Check that the camcorder is powered up (green or red LED indicator) and remove the
lens cap. An image of the photo print should appear on the monitor screen. Move the
camera up and down on its stand until the rock pile fills the field of view. Adjust the
lamps if one part of the image is brighter than another, and to avoid glare. Depending
on the quality of ambient room lighting, sometimes a superior image quality is obtained
with the lamps switched off.
Most camcorders have self-adjusting focus and aperture. If the workstation is equipped
with a camera with fewer automatic features, adjust the aperture by rotating the lower
ring on the lens until the image is neither too dark nor too light, and adjust the focus to
sharpen the image.
For additional information on the installation of WipFrag and its drivers, please refer to
Appendix 13 – WDM Drivers.
8
3. WIPFRAG OPERATIONS
3.1 WipFrag Window
9
3.2 WipFrag Operations
3.3.1 Basic Operation

A typical WipFrag operation will consist of the following:
1. Capturing a new image or open a disk image.
2. Set the scale on the image.
3. Activate the automatic edge detection algorithms.
4. (Optionally) Use the editing tools to improve the edge

detection.
5. Do a virtual sieve, to measure block sizes and generate a graph.
3.3.2 Basic Setup

Some typical WipFrag setup operations will consist of the following operations:
1. Options / Edge Detection Parameters. Set up the parameters that govern the
block edge detection algorithms
2. Options / Sieve Options. Set up titles, models, filename, size classes, calibrations,
and fines sizes.
3. Options / Edit Exclusion Zone. Sets up zones on the image, which are to be
excluded from the analysis.
4. Options / Metric/Imperial. Selects units of measurement, meters or inches. This

can be done in the Sieve Options dialog. You can set the scale units in the Set
Scale dialog.
10
4. WIPFRAG ACTION MENUS
4.1 Main Menus and Toolbars
4.1.1 Opening Menu

The opening menu comes up when there is
no active image.
4.1.2 Image View Menu

The main menu when there is an
active image.
4.1.3 Image View Tool Bar
Only the first, second,

fifth and sixth icons are active when there is no image present.
11
4.2 File Menu and Tools
4.2.1 Open
Opens a bitmap (.bmp) or jpeg (.jpg) file of any size or shape, or a data file (.dat,
.jnt) for graphing. A bitmap file must be an 8-bit grayscale image. In either case,
images should have square pixels.
4.2.2 Open Multiple

Opens one or more data files (.dat) for graphing. If there is more than one data file,
a merged analysis is done. Multiple data file must reside in the same directory.
4.2.3 Capture New Image

The yellow button opens the capture new image dialog box; the green captures
a new image directly.
12
4.2.3.1 Capture new image dialog
Preview - Displays a “live” image.

Snap - Captures an image and transfers it to the host buffer, and exits from the dialog.
Cancel - Exits out of the dialog without capturing and image.
Brightness – Sets the gain of the amplifier on the digitizing board.
Contrast - Sets the offset of the amplifier on the digitizing board.
Defaults - Sets the brightness and contrast to their default values.
Auto-Amp – This feature will try and choose the optimal brightness and contrast for
analysis.
Camera - Selects which camera will be active (1,2,3 or 4) for DFG digitizing board.
Remove Interlace – Removes the video interlace from moving video images by
discarding every other line and interpolating between lines.
4.2.4 Close
Closes an active image.
13
4.2.5 Save
Saves an active image as a bitmap file.
4.2.6 Save As
Brings up the standard Windows Saves As dialog box.
All WipFrag outputs can be saved as JPG, BMP, GIF, EMF, WMF, PDF , WipFrag chart
txt or WipFrag chart dat file.
It can save images as BMP or JPG.
NOTE:
WipFrag saves the current scaling factor with a bitmap image. If WipFrag reads the file
at a later time, it will retrieve the scaling factor for that image.
4.2.7 Open Net

Opens a previously saved network image.
4.2.8 Save Net

Saves an active network image.
4.2.9 Print
Brings up the standard Windows Print dialog box.
4.2.10 Print Preview

Brings up the standard Windows Print Preview.
4.2.11 Print Setup

Brings up the standard Windows Print Setup dialog box.
4.2.12 (Recently used file list)

Allows one of the last 4 .bmp or jpeg files to be loaded, or .dat files to be graphed.
4.2.13 Exit
WipFrag termination.
14
4.3 Edit Menu and Tools
While editing, F-Keys and icons can both be used to change between editing features.
Notes:
1. WipFrag allows manual editing to insert missing boundaries between fused

fragments, and to delete false edges where a fragment has disintegrated into two or
more pieces. In the extreme, editing could be used to trace a complete fragment net
visible to the eye but not to the computer. However, experience indicates that a few
minutes of cleanup is all that is needed in most cases. Editing should start with the
largest fragments that have the greatest impact on weight percentages.
2. The errors of edge detection tend to be systematic and when manual editing is
impractical (for example in on line monitoring operation, these errors can be
compensated, along with corrections for unresolved fines and fragment overlap,
using calibration procedures described in Appendix 4.
4.3.1 Undo
Undo last edit.
4.3.2 Draw Line – F2

Sets the line mode, so that a line is drawn from the point where the left mouse
button is pressed to the point where the mouse button is released.
4.3.3 Draw Polyline – F3

Sets the poly line mode, so that a polygonal line is drawn between points identified
by a left click, terminated by a double left click.
15
4.3.4 Block Delete – F5

Sets the erase box mode, so that all the blues lines in a rectangular box, corners
defined by the point
4.3.5 Delete Line – F4

Sets the delete line mode, so that right clicking on a blue overlay line deletes that
line where the left mouse button is pressed, and the point where the button is released.
All blue lines are erased back to their roots at a red node.
4.3.6 Snap to
Set the mode that blue lines and polylines snap to existing red nodes and green
end points.
4.3.7 Ignore Block – F6

Choose this mode to quickly mark or unmark areas that should be ignored. This
feature is also useful for checking for holes in the net when editing.
4.3.8 Mark as Fines – F7

Choose this mode to quickly mark areas as ‘Fines’ You can allocate a size for fines
in the Options – Sieve Options menu.
16
4.4 View Menu and Tools
4.4.1 Edit Net

Toggles mode to display/hide the overlay net of block outlines.
4.4.2 Edit Exclusion Zone

Brings up the Exclusion Zone menu.
4.4.2.1 Edit Exclusion Zone Toolbar
Camera - Selector for camera number, each has separate saved exclusion zone.
Enable – Enable/Disable action of exclusion zone.
Save - Save exclusion zone.
Erase – Set erase mode; subsequent actions will remove parts of exclusion zone.
Fill - Set fill mode; subsequent actions will add parts to the exclusion zone.
Draw - Fill entire screen with red fill or erase red fill depending on mode.
Rectangle - Carve a rectangle out of or into the exclusion zone.
Polygon - Carve a polygon out of or into the exclusion zone.
4.4.3 Zoom In
Zoom in on the image.
4.4.4 Zoom Out

Zoom out on the image.
4.4.5 Status Bar

Toggle to view status bar.
17
4.5 Fragmentation Menu
4.5.1 Set Scale

Opens the Set Scale Dialog Box.
4.5.1.1 Set Scale Dialog Box
Camera - Selector for camera number, each has separate saved scale.
OK - Accepts the information in the dialog box.
Cancel - Rejects and cancels the information in the dialog box.
Edit Scale - Allows the user to manually enter a scaling factor in pixel per metre.
Calculate – instructions for highlighting the scale bar.
Length - Used to enter the actual length of the scaling object
Pixels – Displays the calculated scaling factor in pixels per metre.
Regardless if the scale is calculated using imperial or metric, the scaling factor is always
in pixels/meter.
NOTES:
1. The scale is saved and is automatically applied to the next digitized image, unless a
new scale is re-entered.
18
2. Selecting a camera number makes it the “active” camera, the scale of which will be
used for subsequent calculations.
3. Saving a .bmp file will save the scale with that file. Consequently a file that has
previous been created by WipFrag will also import the saved scale. However, it will
be stored in the “camera 0” position, irrespective of which camera it was initially
obtained.
4. In no-tilt scaling, the scale is applied equally over the image, and distortions are not
corrected. The longer the scaling object, the more accurate the scaling becomes.
5. The Scaling Factor can be entered manually. This is for situations like conveyor
belts where the scaling is lost, and it would be inconvenient to re-image a scale bar.
Note that the scaling factor cannot be calculated! It must be recorded from a
previous set scale operation.
6. Manually editing the scale can only be done when the scale has been predetermined
– ie. Fixed camera location.
4.5.2 Set Tilt

Opens the Set Tilt Dialog Box for tilt scaling. Tilt scaling is ideal for greatly sloped
muckpiles and other locations where the images cannot be taken perpendicularly to the
material. WipFrag will use this information to properly correct for scale & slope. Please
be advised that this feature must be turned off after use as the last calculated
slope and intercept values will be used for image analysis if left on.
The bottom bar is responsible for the actual scale. The top bar allows for the
calculation of slope.
19
4.5.2.1 Set Tilt Dialog Box
Camera - Selector for camera number, each has separate saved tilt scale.
Snap – Refreshes the live image.
Use Tilt Scaling - Toggle to select tilt or normal scaling.
Upper Scale - Allows the user to manually enter the length of the upper scale.
Lower Scale - Allows the user to manually enter the length of the lower scale.
Slope - Displays the slope of the tilt scaling function.
Intercept - Displays the intercept of the scaling function.
4.5.3 Generate Net

Generate the overlay network of block outlines.
NOTES:
1. The automatic edge detection automatically runs through a series of edge detection
operations to obtain a net of lines corresponding closely to the fragment boundaries.
20
To complete the process typically takes a few seconds. The amount of time to
calculate the net depends on the speed of the computer and the size of the image
being analyzed. The resulting net consists of line segments in blue, line endpoints in
green, and intersections (nodes) in red. This allows you to examine the fidelity of
edge detection; i.e. to check how closely the net matches the boundaries of the rock
fragments. The net edit net and zoom features can assist the comparison.
2. Fidelity depends on the quality of image (lighting and other factors, see Appendix 2),
on the edge detection methods themselves, and on an appropriate choice of the six
edge detection parameters (Section 5.1). If necessary the fidelity can be improved
by adjustments to lighting and camera settings, or by fine-tuning the edge detection
settings, or by manual editing of the net (Section 4.3).
4.5.4 Sieve
Measure the block sizes, generate a disk file and draw a graph.
NOTES:
1. Sieve measures the areas and shapes of net elements. A net must be present,
having been completed by the ‘generate net’ operation, and a scaling factor must
have been entered. Selecting Measure Blocks increments the file identifier
(Appendix 5), and then, WipFrag fills the net elements with pixels in random colors,
one color per element, and counts pixels, converting the counts to areas. The
number of blocks (net elements) is displayed in the message box. The individual
block areas are stored automatically in a DAT file.
2. The area distribution (DAT file) is converted to a volume and weight distribution only
at the graph output stage. The computed fragmentation statistics can be stored in a
LOG file. Selected data are displayed and printed in the graphs. The 2D to 3D
conversion uses a mathematical unfolding function.
21
4.6 WipJoint Menu
NOTES:
1. WipJoint is an optional model that must be specifically enabled on the security

key; otherwise the WipJoint operations cannot be used.
2. WipJoint uses dual edge detection operators, one to capture the linear elements
of the structure for orientation and the other to force closed polygons around
blocks for apparent block sizes. Both nets need to be generated in order to do
the analysis.
4.6.1 Set Scale

Opens the Set Scale Dialog Box .
4.6.2 Generate Edges

Generate the network of structure linear edges.
4.6.3 Generate Polygons

Generate the network of block outlines.
4.6.4 Edit Edges

Enable editing of structure linear edges.
4.6.5 Edit Polygons

Enable editing of block outline polygons.
NOTES:
1. The resulting net consists of line segments in blue, line endpoints in green, and
intersections (nodes) in red. This allows you to examine the fidelity of edge
detection; i.e. to check how closely the net matches the boundaries of the rock
fragments. The edit net and zoom features (Sections 4.4.1, 4.4.3, 4.4.4) can
assist the comparison.
22
4.6.6 Analyze Joints

Analyze network of highwall structure outlines and polygonal blocks.
23
4.7 Right Click On Image
4.7.1 Zoom In
Zoom in on the image.
4.7.2 Zoom Out

Zoom out from the image.
Note: You can also use the mouse wheel to zoom in and out.
4.7.3 Ignore This Block – F6

Block gets marked so that it is ignored in the sieve analysis.
4.7.4 Mark As Fines – F7

Block gets marked so that it is interpreted as a zone of fines in the sieve analysis.
24
4.10. OPTIONS MENUS
4.10.1 Edge Detection Parameters
Camera - Selector for camera number, each has separate saved edge detection
parameters.
Save – Refreshes the current set of parameters.
Restore - Restores a saved set of parameters.
Tom’s Picks – A set of parameters recommended by Tom.
Presets 1-9 - Nine preset combinations of edge detection parameters.
Window Size - see section 5.1.1.
Threshold - see section 5.1.1.
Valley Threshold - see section 5.1.1.
Search Length 1 - see section 5.1.1.
25
4.10.1.1 Edge Detection Settings

Edge detection parameters are numerical values used by WipFrag during the various
stages of fragment edge detection. The default settings shown in the above menu are
recommended for normal operation with good image quality. Judge the results by
comparing the net with the image.
If net fidelity appears less than ideal, you can try replacing the default settings by one of
nine alternative preset combinations of values, by repeated clicking on Preset. Preset 1
gives the most edges and Preset 9 the least. (Try Preset 5 as a starting point). WipFrag
will use the selected values until you make further changes. Check the net-on-rock
overlay to see whether edge detection is improved. Select WipFrag Defaults to restore
the default values at any time.
As a last resort for images of poor quality you can individually adjust each parameter by
clicking on the field and entering a new value within the permitted range. However,
individual adjustments often make matters worse and should be avoided by new users.
Individual combinations can be saved using the Save settings option, and recalled later
using the Restore settings options.
Again, you can select Old WipFrag Defaults to restore the original default values at any
time.
Tom’s picks are a set of parameters that seem to work well with heterogeneous rock
assemblages like blast fragmentation.
The following is a description of the parameters:
Window Size is the length in pixels of the side of a square window used for
thresholding. The image size is 640 by 480 pixels (768 by 574 for PAL versions) and
the largest fragment should measure about 100-130 pixels across, about 15% to
20% of the image width). Decreasing the Window Size will tend to cause fusion of
finer particles, whereas increasing it leads to breakdown of larger particles and loss
of information in shadows.
Threshold is the difference in intensity (gray tone level, range between a pixel and its
window average. Increasing the threshold results in fewer blocks (some blocks
falsely eliminated), and decreasing the threshold results in more blocks (some
falsely identified).
Valley Threshold specifies a minimum level of gray tone slope to trigger. Increasing
the Valley Threshold gives fewer blocks (some falsely joined together) and
decreasing it produces more blocks (some of the larger ones disintegrate).
26
Search Lengths 1, 2 and 3 are the radial search lengths in pixels measured from a
given vertex that the segment operator searches to find and join. Search Length 1
should be greater than Search Length 2, which should be greater than Search
Length 3. Increasing these parameters results in more blocks (some falsely
identified), whereas decreasing them results in fewer blocks (some blocks falsely
eliminated).
4.10.1.2 Strategy
1. Try starting out with preset 5.
2. Change the window size so that it is the same size in pixel as the largest
common block. (The size of a block in pixels can be determined by putting the
mouse cursor first on one side of the block and then on the other, and viewing
the difference in the x and y coordinates given in the bottom right hand corner of
the screen).
3. Temporarily set the 3 search lengths to the minimum permissible values.
4. By trial and error (simultaneously) determine the best value for threshold and
valley threshold.
5. Increase the search length settings. As these are increased, first the number of
missing edges will decrease, but later the number of false edges will increase.
Pick an optimum value.
NOTES:
1. For a given application, i.e. same camera, images taken at the same scale of
observation, under the same lighting conditions, the edge detection parameters
should stay constant.
2. For any comparative study, edge detection parameters, as well as scale of
observation, lighting conditions, camera etc., must remain constant.
4.10.1.3
Edge Detection Parameter Default Setting Tom’s Pick Preset 5 Range
Window Size 50 125 70 30 to 150

Threshold -9 -5 5 -15 to +25
Valley Threshold -1.0 -2.0 -2.0 -4.0 to -0
Search Length 1 24 15 25 5 to 45
27
4.11 Canny Parameters
Camera - Selector for camera number, each has separate saved edge detection
parameters.
Save – Refreshes the current set of parameters.
Restore - Restores a saved set of parameters.
Tom’s Picks – A set of parameters recommended by Tom.
Lower Threshold – Canny threshold for beginning tracking.
High Threshold – Canny threshold for ending tracking.
Gaussian Standard Deviation – Standard deviation of Gaussian smoothing function.
28
4.11.1 Canny Edge Detection Settings

Edge detection parameters are numerical values used by WipJoint during the generate
edges stage. These are selected to produce the best possible linear features on the
image on which the orientation of joints are measured.
The following is a description of the parameters:
Gaussian Standard Deviation is the strength of the Gaussian smoothing convolution.

Decreasing the Standard Deviation will result in more real (and false) edges being
identified. Where as increasing it will result in less real (and false) edges. The
analysis is very sensitive to this parameter.
Lower Threshold is the magnitude (intensity) on the gradient image at which point
tracking begins. The analysis is relatively insensitive to the value of this parameter.
Upper Threshold is the magnitude (intensity) on the gradient image at which point
tracking ends. Decreasing the upper threshold will result in less edges (and less
noise) being identified. The analysis is relatively insensitive to the value of this
parameter.
4.11.1.1 Strategy
1. Try starting out with Gaussian Standard Deviation of 1.
2. If there is too much noise in the image, increase this value.
3. If there are too few edges in the image, decrease this value.
4. Usually, the upper tracking threshold can be set quite high and the lower
threshold quite low for good results. Setting the lower threshold too high will
cause noisy edges to break up. Setting the upper threshold too low increases the
number of spurious and undesirable edge fragments appearing in the output.
29
4.12 Sieve Options
Camera - Selector for camera number, each has some separate saved options.
Filename Prefix - User defined 4-character prefix for data file names.
Distribution Model - Selects model, use well graded for any type of fragmentation
analysis.
Size Classes - Toggles between ISO and US Sieve sizes for output translation of the
chart.
Titles - Two layers of titles for the graphical output.
Log to File - Option to output analysis information to a csv log file.
Filename – User specified name of the log file.
n (Calib. Target) - Calibration target Rosin-Rammler n-value.
B (Curve Undulation) – Calibration Factor for Swebrec
Rock Density - Density of the rock in kg/m3.
n Adjust - Multiplicative adjustment factor for Rosin-Rammler n-value.
Xc Adjust - Multiplicative adjustment factor for Rosin-Rammler Xc-value.
Xmax Adjust – Correction for Swebrec Xmax value.
X50 Adjust – Correction for Swebre X50 value.
Average Fines Size – Size used for areas demarked as “fines” areas.
30
NOTES:
1. Log to File option – the header written on the csv worksheet will be in the units
selected when the sheet was first created. If the output measurement is changed
(metric or US Sieve) the values will remain accurate, but with the wrong header.
Ie. It will say 5.68 under the column of D90 – where the header is meters but the
value written is in inches. Please be advised, that if the output is changed, a
new csv ‘Log’ file must be created.
2. The user-supplied filename can be up to 4 characters long and is supplemented
by a WipFrag generated sequence number to uniquely identify each image and
each analysis. More details about file naming conventions can be found in
Appendix 5.
3. The maximum length of a title line is 40 characters.
4. Average Fines Size is used to enter the average size of fines applied to a
manually identified zone of fines.
5. Output Log File stores in a single file the statistics generated by successive
WipFrag runs. Each new run appends statistics to the currently selected LOG
file. For this option to be activated, the file must first be named. Select Log File
and enter a filename. To terminate logging, click on this option a second time and
the filename will be replaced by none. Restart a new log file by entering a new
name.
6. Rock Density (kg/m3) is an option used in converting 2D areas to 3D volumes.
Note that the rock density is has an effect on the log histogram only.
Normalization In the case of the cumulative graph removes the effect of rock
density. Rock Density expects a value of density appropriate for the rock in the
image (as tabulated below). This value is used to convert from volume to mass
and weight. The default value of 2670 kg/m3 is realistic for a typical dense quartz-
rich igneous rock such as granite or a low-porosity limestone or sandstone. Dark
colored mine rocks and ores may have densities of 3000 or more.
Typical rock densities (kg/m3)

(1 g/cm3 = 1000 kg/m3 = 1 t/m3 = 62.423 lb/ft3)
Type Min. Max.
Shale 2170 (soft) 2520 (hard)
Sandstone 2290 (porous) 2700 (dense)
Limestone 2440 (porous) 2720 (dense)
Granite 2670
Norite 2980
Dunite 3280
31
7. Distribution Model selects between two separate analytical models for the 2
dimensional to three-dimensional unfolding function. The “well graded” model is
to be used for blast fragmentation assessment whereas the “highly uniform”
model is to be used for applications such as crusher products on conveyor belts.
8. Calib. Target is a calibration selector based on the expected Rosin-Rammler n
value of the distribution being measured NAdjust and XcAdjust are adjustment
factors that are used by multiplying them with the measured Rosin-Rammler n
and Xc values, and generating a dashed Rosin-Rammler cumulative curve.
9. WipFrag converts the measurements of 2-D net segment areas into 3-D fragment
volumes and weights using a mathematical unfolding function together with an
empirical correction to account for fragment overlap, shape effects and missing
fines (i.e. particles too small to be resolved in a single full-scale image). Empirical
correction factors Fn and Fxc rotate and translate a Rosin-Rammler curve fitted to
the WipFrag data in log-log co-ordinates. The R-R parameters n and Xc for the
raw data are multiplied by NAdjust and XcAdjust to obtain a close approximation
to the "true" values that would be measured by sieving a crushed rock calibration
standard, or by a zoom-merge procedure.
10. Values for the correction factors depend to some extent on to the size-uniformity
of the rock pile, given by the Uniformity Coefficient "n", equal to the slope of the
gradation curve. This is itself a parameter to be measured. Choose an initial
estimate of n to select an appropriate pair of correction factors from the above
table. Repeated clicks on n (target value) allow you to toggle between the six
uniformity conditions and select the condition that most closely represents the
uniformity of your typical image (if in doubt, select n = 1.0). If the measured n
shown on the output graph is different, select a new target value and repeat the
run (iterate) until the target and measured n values are similar.
11. Select None to process the results of a calibration. This will set the correction
factors to 1.0 to provide "raw" (uncorrected) values for n and Xc. Divide true
values (determined by sieving or by zoom-merge) by the raw values to obtain a
new set of correction factors.
12. The following correction factors have been determined from experiments using
sieved samples, and the well graded model:
32
13.
Uniformity (True n) NAdjust XcAdjust

n = 0.50 very non-uniform 0.30 0.50
n = 0.75 lower limit of blast fragmentation 0.45 0.90
n = 1.00 non-uniform 0.48 1.00
n = 1.25 average uniformity 0.60 1.00
n = 1.50 upper limit of blast fragmentation 0.65 1.00
n = 2.00 uniform 0.78 1.00
n = 3.00 highly uniform 1.15 1.00
See Appendix 3 for an explanation of how to prepare correction factors for your own
rock samples.
Please note that these calibrations have been derived for the “well graded”
distribution model. These should not be used with the “highly uniform” distribution
model.
33
4.12 Window Menu
4.12.1 New Window

Select this command to create a new window identical to the active window.
4.12.2 Cascade
Select this command to arrange the open windows so they cascading one over the
other.
4.12.3 Tile
Select this command to arrange the open windows as tiles within the main frame
window.
4.12.4 Arrange Icons

Arranges the icons of the windows you have minimized, neatly at the bottom left of the
main frame window.
34
4.13. Help Menu
4.13.1 Help Topics

Select this command to launch the WipFrag help application. WipFrag help contains
information about each menu and dialog box in WipFrag.
4.13.2 About WipFrag.

Select this command to display the About WipFrag dialog box. This dialog box displays
information about the version number of WipFrag, data read from the sentinel security
key, the type and description of the frame grabber found on your computer, TTL I/O.
Also links to WipWare’ website and e-mail address are available.
35
APPENDIX 1: WIPFRAG “FINES” STRATEGIES
WipFrag uses geometric probability theory to unfold a 3-D distribution. This accounts
for partially overlapped fragments, as well as missing fines to some extent. However,
since the number of fines in a distribution, which are too small to be identified on the
image, are highly variable, impossible to measure, and difficult to predict, one of the
following solutions is suggested:
Engineering Solution
The engineering solution provides a quick answer with a minimal amount of effort. This
involves taking one or more images and simply analyzing them.
This solution works well for comparative purposes and for narrow (poorly graded)
distributions. For wide distributions (well graded) it must be recognized that the missing
fines will result in a measurement bias toward the larger sizes.
Empirically Calibrated Solution

Like the engineering solution, the empirically calibrated solution provides a quick answer
while doing production measurement. Again, this involves taking one or more images
and analyzing them using the empirically derived Rosin-Rammler (R-R) correction.
The appropriate calibration factor, (factors determined by laboratory sieving) is selected

estimating the slope of the R-R curve of the actual distribution. Where appropriate,
calibration factors can be determined by a full scale sieving trial by the user under the
conditions and rock types appropriate to the measurement.
This method is much more accurate than the engineering solution. Analysis is quick
and simple, but a one-time investment must be made for calibration, prior to the start of
analysis.
Zoom-Merge Solution
Unlike the previous two solutions, this solution provides a measurement that is as
accurate as possible, and should be employed when a more time consuming analysis
can be tolerated.
This, the most accurate of solutions requires multiple images to be analyzed, at different
scales of observation. It is time consuming, as it requires significant manual interaction,
and organization of samples and files.
36
APPENDIX 2: ROCKPILE SAMPLING AND PHOTOGRAPHY
Rock pile Sampling

You can only measure what you can see; so reliable results call for some basic
precautions in selecting suitable viewpoints, and in photography.
Develop a sampling strategy. Taking images is like taking samples for sieving or
other tests. The results must statistically represent the product as a whole.
Take several shots, preferably at least five at random locations on a large rock pile,
or of several truckloads or draw points. WipFrag allows the results for several
images to be merged and averaged as a single data sample.
Include all sizes. No single block should occupy more than 20% of the width of
image. For improved resolution of fines, use WipFrag's zoom-merge capability to
combine images at different scales of magnification. For improved estimates of
oversize, increase the number of full-scale shots to at least ten.
Beware of rock pile segregation. Large blocks tend to roll to the outer edges and
fines may cover the surface or become hidden as a result of gravity or rainfall. The
effects can be minimized by increasing the number of images per sample but only
with careful selection of image locations. For example, take slices through the pile,
or photograph the rock during loading or in trucks. However, segregation usually
increases with digging and other forms of materials handling.
Photography and Lighting

Fill the field of view with fragmented rock, preferably at least 400 fragments and up
to about 2000 fragments.
Include at least one scaling object of known length, such as a 2 m white scale bar
along the lower edge of the image. Alternatives include beach balls, hoops, parallel
laser beams, or the known widths of truck or conveyor.
Position the scale near the edge of the image so as not to obscure the rock you are
trying to measure. When using a video camera at a fixed object distance, the scale
is required only during the initial or final frames.
Avoid wide-angle close-up photography and oblique shots that distort the scale. If
the rock pile surface is oblique to the camera, place identical scaling objects at the
nearest and furthest points that can be averaged or used in auto-tilt correction.
Provide uniform indirect or diffuse lighting without excessive sharp or one-sided

shadows and "hot spots". Video cameras do not require brilliant illumination.
37
WipFrag works best when each fragment is equally bright and surrounded by a thin,
uniform shadow.
For underground photography, choose area illumination with several medium-power

lamps in preference to direct flash or spotlight. Copy stand photoanalysis of prints or
crushed rock samples is often best with ambient fluorescent or window illumination.
For outdoor work, choose dull days in preference to bright sunlight.
Maintain cameras and lamps in good working condition. Those mounted at fixed
locations will require protection from dust and mechanical damage, in some cases
from blast damage. Clean lenses using compressed air to avoid scratching.
Note: After years of experience, we have found that acquiring images with a Video
camera can be greatly beneficial, as audible notes can be made while taking the
footage. Location, scales used and any other conditions can easily be recorded while
video recording the material.
38
APPENDIX 3: CALIBRATION SAMPLES AND TECHNIQUES
WipFrag inherently converts the measurements of 2-D net segment areas into 3-D
fragment volumes and weights using a mathematical unfolding function. This accounts
for fragment overlap, shape effects and missing fines (i.e. particles too small to be seen
in a full-scale image) to some extent.
The optional empirical correction in WipFrag uses factors Fn and Fxc to rotate and
translate a Rosin-Rammler straight line fitted to the WipFrag data in log-log co-
ordinates. The R-R parameters n and Xc for the raw data are multiplied by Fn and Fxc to
obtain the "true" values that would be measured by sieving a muck pile or a crushed
rock calibration standard, or by a zoom-merge procedure. The values of Fn and Fxc will
range between 3.0 for good images with narrow size distributions (well sorted), to as
low as 0.5 for poorer images with wide size distributions (well graded). WipFrag default
values for the correction factors have been obtained for n = 0.5, 0.75, 1.00, 1.25, 1.5,
2.0 and 3.0, but other values may be more representative of local blasting.
To calibrate the WipFrag system, sieving of a muck pile, after analyzing with WipFrag
results in the best values. However because of the effort required, alternatively scaled
down laboratory crushed rock calibration samples with linear R-R size distribution and
uniformity coefficients covering the full range anticipated in the mine can be used. The
samples should have similar color, sphericity and general appearance to the typical rock
to be analyzed.
Crush sufficient rock, about twice the total quantity required per sample. Separate into
size fractions using a nest of sieves of different mesh sizes. Use an automatic shaker
for ten minutes. Weigh and record the weight fractions to determine the true Xc and n
values.
Re-mix the sample, and run analyses about ten times per sample, shaking and re-
mixing the sample between each run. Merge the data and obtain uncorrected Xc and n.
Divide true values by raw values to obtain a new set of correction factors. Repeat for
other n values. Compare the results with the defaults in WipFrag and if required, make
changes to the calibration settings.
39
To make changes to the calibration edit the calib.cnf file in the WipFrag directory. The
file should look something like this:
0.5 0.30 0.50 1.75 0.75 0.75

0.75 0.45 0.90 2.00 0.75 1.00
1 0.48 1.00 2.15 0.75 1.25
1.25 0.60 1.00 2.30 1.00 0.75
1.5 0.65 1.00 2.45 1.00 1.00
2 0.78 1.00 2.60 1.00 1.25
3 1.15 1.00 2.75 1.25 0.75
3 1.15 1.00 2.90 1.25 1.00
3 1.15 1.00 3.25 1.25 1.25
There are six columns: N, N Adjust, Xc Adjust, b, Xmax Adjust, and X50 Adjust. Every row needs to
contain all six values. The first three values are for R-R calibration while b, Xmax and X50 adjust are for
Swebrec Corrections
Please note that these calibrations have been derived for the “well graded” distribution
model and are to be applied to blast fragmentation only, and not to be used with the
“highly uniform” model (Section 7.3). Using the above calibration techniques requires
the estimation of the true or target n-value of the distribution. This can be a visual
estimate. While the human eyes are not very good at estimating sizes, it is rather better
at estimating the variability of the distribution.
The suggested method is to use the picture of the type distributions as a comparative
tool to estimate n-value. These pictures can be found in the accompanying paper
entitled “Calibration of optical digital fragmentation measuring systems”.
40
APPENDIX 4: DEFINITIONS OF WIPFRAG STATISTICS
Granulometry is the measurement of particle (grain, fragment) sizes and

shapes. The science of granulometry, only recently applied to blast fragmentation, has
been developed to an advanced stage in earlier studies of comminution (crushing and
grinding), petrology (study of rocks and minerals under the microscope), soil mechanics
(classification of soils), metallography, powder technology, and other fields. A confusing
range of alternative definitions exists even for such fundamental items as fragment size,
which can be measured in terms of projected lengths, diameters of circles with the
same area, and in several other ways, and the choice becomes even more complex
when selecting statistics to represent fragment size distributions and fragment shapes.
The following is a definition of some granulometry statistics selected and employed in

WipFrag.
WipFrag Statistical Values:
Cu Coefficient of Uniformity = D60/D10, a measure of the slope of the

cumulative weight % curve between the 60- and 10-percentiles.
Dn Nominal diameter, or equivalent spherical diameter, i.e. the diameter of a

sphere with the same volume as that computed for the fragment.
D10, D25 etc.
Percentile sizes. For example D10 is the ten-percentile, the value of De for
which 10% by weight of the sample is finer and 90% coarser. In terms of
sieving, D10 is the size of sieve opening through which 10% by weight of
the sample would pass.
D50 The Median or 50-percentile, the value of Dn for which half the sample
weight is finer and half coarser.
Blocks Number of net elements detected in the NET image
Max Maximum size of fragment in the image [Dn (m)]
Mean Arithmetic mean (average) fragment size, equal to the sum of all
equivalent spherical diameters divided by the total number of particles
[Dav (m)]
Min Minimum size of fragment in the image [Dn (m)]
Mode Most common sized particle, the geometric mean Dn size class interval for
the class containing the greatest number of net elements (fragments) [Dn
(m)]
41
N Rosin-Rammler (and Gaudin-Schuhman) Uniformity Coefficient, equal to

the slope of the Rosin-Rammler straight line fitted to the data in log-log co-
ordinates.
Sphericity Dn/Ds, the ratio of equivalent spherical diameter to the diameter of a

circumscribing sphere (long axis of the fragment)
stdev Standard deviation of fragment size Dav
Xc Characteristic Size, the intercept of the Rosin-Rammler straight line fitted

to the WipFrag Dn data in log-log co-ordinates. This is equivalent to the
D63.2.
Xmax Gaudin-Schuhman characteristic size, the intercept of the 100 % passing

and the slope of the Gaudin-Schuman straight line.
WipFrag Output Values:
Xmax2 Is the calibrated Gaudin-Schuhman characteristic size, the intercept

of the 100 % passing and the slope of the Gaudin-Schuman straight line.
Xc2 Is the calibrated Characteristic Size, the intercept of the Rosin-Rammler

straight line fitted to the WipFrag Dn data in log-log co-ordinates. This is
equivalent to the D63.2.
Nros2 Is the calibrated Rosin-Rammler (and Gaudin-Schuhman) Uniformity

Coefficient, equal to the slope of the Rosin-Rammler straight line fitted to
the data in log-log co-ordinates.
WipJoint Definitions:
The Spacing Rosette is generated by measuring the distance between joints on the
rock face. WipFrag measures these spaces in 18 direction classes of 10 degrees each.
If the spacing is wide in the horizontal direction, and narrow in the vertical direction, the
rosette will reflect this by being wide and flat
Max = Is the average spacing in the class with the highest spacing
Min = Is the average spacing in the class with the lowest spacing
Ratio = Is simply the max/min. In a sense this characterizes the aspect ratio of the
"typical" block.
42
The Orientation Rosette is generated by measuring the length of joints for each of 18
direction classes (10 degrees each) on the rock face. This rosette will illustrate in which
plane the jointing is predominant.
Max = Is the total length of all joints in the principal joint plane.
43
APPENDIX 5: FILE NAMING AND GRAPH LABELING CONVENTIONS
WipFrag file names are composed of a user ID file label XXXX entered via the names
and titles submenu, an image identifier YY incremented automatically by computer each
time an image is grabbed, and an analysis identifier Z incremented whenever blocks are
measured and counted:
XXXX_YYZ
Because all alphanumeric characters are used, this gives unique labeling for up to 1296
(36x36) images per user-supplied name.
Data files are saved automatically with extension .DAT whenever blocks are measured
and counted (.JNT files for WipJoint). Image files with extension .ROK, .NET, .BMP, are
saved only when specified by the user either using the current filename or a user-
specified filename. Log files are saved automatically with the user-specified name plus
.LOG. Log files consist of a single row of data and are continually appended for each
analysis, until the user specifies that a different filename be used.
WipFrag also has four titles (captions) for graphical output and LOG and DAT files, two
user-supplied, and two computer-generated. For example:
North Bench Blast 24b

Performed by John Doe
WipFrag (c) ver. 2.4 build 41 Tue 24 Jan 2006
WipWare Inc. Copy ID=ver_001
The first two lines supplied by the user via the options submenu. They can be changed
for each analysis or retained for a longer series of analysis. The third line is a WipFrag
generated copyright, date and time stamp combination. The fourth line is a factory
installed user identifier that cannot be altered, together with a current file identifier.
WipFrag stores the current titles, and displays the first 18 characters of each title in the
status box. Current titles and time stamps are written into the data files; the stored titles
and date/time are printed on the graphs. When data files are merged to produce graphs,
because they each might have different titles and date/time stamps, the current title and
current date/time stamp are used instead.
44
APPENDIX 6: WIPFRAG SECURITY KEY
WipFrag Software requires the use of a security key and comes with a limited license or
an unlimited license. In addition, WipJoint, an optional add-on requires the security key
to be programmed to run WipJoint.
Unlimited License
This version simply requires the presence of a Sentinel Security Key provided with the
WipFrag software. At startup, and at various times during execution, the WipFrag will
check for the presence of the security key.
The software may be copied onto more than one computer, however because of the
need for the physical presence of the key, WipFrag can only be run on one computer at
a time.
Restricted License
Under some circumstances, versions with restricted licenses will be released. This
security is built into the Sentinel Security key. Time can be limited. A word of caution:
Changing system time to daylight savings time, or putting the key on a second computer
with a different system time could inadvertently cause WipFrag to shut down.
The security key logs the use of WipFrag, and if the system time on the computer is
accidentally changed with (i.e. set back) WipFrag will shut down, and the key will have
to be sent back to WipWare for re-programming. Reprogramming, shipping and
handling charges will apply.
Licensing Changes
Upgrading a restricted license to more uses or a later expiry date requires a
replacement key, or requires the key to return to WipWare and re-programmed.
Reprogramming, shipping and handling charges may apply.
Upgrading a restricted license to an unrestricted license can be done with software

shipped on a diskette.
LOST OR DAMAGED SECURITY KEYS

Lost keys cannot be replaced! Please make sure that you do not loose your security
key. WipWare Inc. recommends that you insure your security key for the full value of the
WipFrag software.
Damaged security keys can be returned and replaced at a nominal cost, at the
discretion of WipWare Inc.
45
APPENDIX 7: EXAMPLES OF WIPFRAG OUTPUT
46
APPENDIX 8: IMAGING MOVING PICTURES
When analyzing moving images such as on conveyor belts or falling off the end of
conveyor belts special efforts have to be made to prevent motion blurring of the image.
Primarily, the shutter speed of the camera needs to be high, at least 1/1000 sec., and
possibly up to 1/10,000 sec. depending on the speed of the motion, and the size of the
image. For some cameras, this may require auxiliary lighting, either fixed or strobed.
Interlaced Video
Standard video cameras use interlaced images based on television standards. That is
to say, a single picture, or frame (NTSC) which has approximately 480 lines is
composed of two separate fields of 240 lines each. These two fields are captured by
the camera with 1/60 second spacing between fields.
On a television monitor, the two fields are updated alternatively, and this interlacing is
all but imperceptible. A video framegrabber on the other hand combines the two fields
into a single frame that is frozen. In this case motion is very much perceptible and
becomes a problem for the subsequent analysis. Despite the fact that a very high
shutter speed (1/1000 second) may have been used, the motion blur is captured in the
1/60 sec. lag between fields.
The video interlace can be removed in software in the options menu. In this case the
analysis uses a single field (640 x 240 pixels for the analysis) rather than the full frame
640 x 480 pixels (NTSC). To retain the correct aspect ratio, the single video field is
duplicated to create a full frame.
47
APPENDIX 9: CAMERA ASSIGNMENTS
Most of the setup parameters are camera specific; i.e. there are separate values for
each parameter based on which camera the image was acquired from. Consequently
parameters must be selected for each camera.
Camera 1, 2, 3 and 4 apply to inputs on the DFG/LC4 card. Camera 0 applies to files
loaded from disk. Using the MRT input cable nominally means camera 1.
Camera Switching
Internally, WipFrag keeps track of which camera is active. The active camera can be
toggled by the user on the various menus on which the camera select item appears.
Automatic switching however takes place under the following circumstances:
When a disk file is loaded, camera 0 is automatically selected.
If camera 0 is active, and a grab is initialed, camera 1 is automatically selected.
Affected Parameters
All edge detection parameters.

Scaling factors. Exclusion zones.Rock density, distribution model, size classes,
calibration model.
Non-Affected Parameters
Titles, filenames, log filenames.
48
APPENDIX 10: OPTIONAL WIPJOINT MODULE
WipJoint
WipJoint is an optional software feature that can be added to the WipFrag®

photoanalysis system that enables the user to characterize and measure jointing
patterns that are apparent on in-situ rock surfaces. In the same way that WipFrag
measures the geometry of blasted rock, WipJoint characterizes the pattern of geological
joints, faults, and other lineations in the in place rock mass. These measurements can
be made at any scale from microscopic to satellite images.
Due to the specialized nature of this measurement and the multitude of variables,
textures, colors, lighting conditions and applications manual input of joints is required for
reliable results.
Jointing Measurement
In engineering applications joints are important not only because they limit the strength
of the rock mass, but also because they control bulk deformation and the flow of
groundwater. At one extreme, the mechanical behavior of an intensely deformed rock
mass approaches that of a granular soil. Large-scale rock mass behavior must be
taken into account in all rock-engineering problems
WipJoint has been developed for use in structural mapping of highwalls in open pits and
development faces in underground mining and construction. WipJoint is used for joint
mapping, which is essential for accurate rock quality designations and blasting designs.
WipJoint measures joint orientation and spacing, which are important considerations in
blasting design, wall control and stabilization, and which feature prominently in most
rock mass classifications. WipJoint will indicate the pre-blast apparent in- situ block size
from the jointing pattern data that is evident on the rock face.
WipJoint can provide a permanent data and image record for archival purposes.
Contractors involved in tunneling will find this valuable in documenting ground
conditions encountered as the tunnel proceeds.
Quarry operators will be able to pre-determine the “blastability of material” based on
jointing data and assess the damage caused by previous blasting and pre-shear efforts
on the highwall. Mine planners will be able to adjust patterns based on jointing
characteristics to minimize overbreak and reduce underbreak resulting in more
competent support and safer openings.
49
WipJoint users will be able to assess the comminution factor which is the mean block
size based on jointing before the blast divided by the mean block size measured after
the blast. This factor can be used to predict fragmentation based on jointing data for
future shots and blast models can be adjusted to produce more accurate results.
The WipJoint Method
WipJoint starts with a video image of a rock face or surface covered by lineations.
The user first sets the scale of the image (Set Scale)
The user then selects Generate Edges and then Generate Polygons to create the
outlines needed (edges for linear elements for orientation measurement, polygons for
closed elements for sizing of blocks.
The user then selects editing mode (Edit Edges, Edit Polygons) to allow manual tracing
of the outlines of the joints on the screen using the mouse, generating a joint network or
“net”. This can be done for both of the above buffers.
The user then selects Analyze Joints. This measures the net and displays and plots a
selection of statistics and graphs.
Joint Measurement
Analyze Joints measures the areas and linear elements of the net. A net must be
present in the window, having completed either a manual editing or a Load Image File
operation, and a scaling factor must have been entered. Selecting Analyze Joints
increments the file identifier, and then:
1. Measures the spacing between joints (in all directions) by traversing the image
with scanlines,
2. Measures the orientation of all linear elements by tracing them automatically,
and,
3. Measures the apparent block size of the closed polygonal element by filling
the elements with pixels in random colors, one color per element, and counting
pixels, converting the counts to areas.
The number of blocks (net elements) is displayed in the message box. All measurement
results are stored automatically in a JNT file.
50
Joint Graphs
WipJoint puts out a composite graph consisting of three parts (See example in
Appendix 13).
The first is a Spacing rosette, giving the mean values of spacing for each of 18 direction
classes of 10 degrees each.
The second is an orientation rosette, giving the total length of joints in each of 18
direction classes of 10 degrees each.
The third is a Log Histogram similar to the one produced by WipFrag giving a non-
cumulative distribution of fragment weights in each of the logarithmically subdivided size
classes or "bins". The linear y-axis scale gives the weight (or volume or mass) of
fragments in each class interval as a percentage of the total "sample" of measured net
elements. The logarithmic x-axis scale gives the fragment size Dn, the nominal diameter
of a sphere with the same volume as the fragment.
51
APPENDIX 11: OPTIONAL WIPJOINT IMAGE SAMPLING

You can only measure what you can see; so reliable results call for some basic
precautions in selecting suitable viewpoints, and in photography.
The image should show a relatively flat cut or outcrop, from an orthogonal viewpoint.
Develop a sampling strategy. Where possible get complete coverage of the rock
face with a single image or a mosaic of multiple images.
Choose a scale of observation that is appropriate, e.g. zoomed in close enough so
that the smallest joints of interest are clearly visible in the image.
Enhancement of Joints on the Face
It is possible to enhance the joint traces on the face and make them more discernible by
one of the following methods:
Low angle lighting will make joint traces associated with relief easier to see.
However low angle lighting will also add an orientational bias to the measurement.
Wetted, and partially dried rock faces tend to make joint traces easier to see, if the
joint are able to absorb the water, as they tend to hold the water longer than the rock
face.
Similarly joint traces can be made more visible if they can be induced to take in
florescent dye, and which can be excited by ultraviolet light.
Photography and Lighting
Fill the field of view with rock, avoiding large expanses of sky or blasted rock rubble.
Avoid pictures with excavators or cranes or other machinery wherever possible.
Include at least one scaling object of known length, such as a horizontal black and
white scale bar along the lower edge of the rock face. If multiple images are used to
cover the face, provide registration marks.
If the orientation of the camera is not close to orthogonal to the rock face, (i.e. in the
vertical direction) use two scaling bars, one at the top of the rock face and one at the
bottom, and use tilt scaling.
Avoid wide-angle close-up photography and oblique shots (i.e. in the horizontal
direction) that distort the scale.
Hold the camera as near to horizontal as possible: measurements of spacing and
orientation are angle dependent. Use a tripod with a bubble level for added
accuracy.
Provide uniform indirect or diffuse lighting without excessive sharp or one-sided
shadows and "hot spots". Video cameras do not require brilliant illumination.
Maintain cameras and lamps in good working condition. Those mounted at fixed
locations will require protection from dust and mechanical damage, in some cases
from blast damage. Clean lenses using compressed air to avoid scratching.
52
APPENDIX 12: OPTIONAL WIPJOINT OUTPUT
53
APPENDIX 13: WDM DRIVERS/V. 2.4 INSTALLATION
Beginning with version 2.4, WipFrag now support WDM (Windows Driver Model) video
capture devices. In theory all capture devices with appropriate WDM drivers are
supported, currently we provide support for the following devices:
1. DFG/LC4 4 camera PCI framegrabber from The Imaging Source.

2. VCE-B5A01 2 camera PCMCIA framegrabber from Imperx.
We will shortly be support DV (digital video) cameras.
Version 2.4 Installation

The following is the procedure for installing
1. Install DirectX. (DirectX 9.0b End-User Runtime) available from Microsoft or on

our Install CD. For Windows Xp this may not be necessary.
2. Install the video capture card and the driver supplied by the manufacturer of the
card.
3. Verify that the video capture card is working correctly using utilities supplied with
the card.
4. Install the WDM drivers from the WipFrag CD.
5. Install WipFrag 2.4 from the WipFrag CD.
6. Run WipFrag.
Running version 2.4 for the first time.
The first time WipFrag is executed, or when the Directx/Properties menu item is
selected, the following video initialization Wizard appears:
54
This assigns the physical video inputs to a camera number in WipFrag. Next, a video
mode needs to be selected as follows for the DFG/LC4:
55
APPENDIX 14: EXAMPLES OF WIPFRAG OUTPUTS
The csv outputs for WipFrag appear as
Year,Month,Day,Hour,Minute,Second,D01,D05,D10,D20,D25,D30,D50,D70,D75,D80,D
90,D95,D99,Xc,Nros,Xmax,0.60 mm,0.85 mm,1.00 mm,1.40 mm,2.00 mm,3.35
mm,4.00 mm,4.75 mm,5.60 mm,6.70 mm,8.00 mm,10.0 mm,12.5 mm,16.0 mm,25.0
mm,31.5 mm,35.5 mm,37.5 mm,40.0 mm,50.0 mm,75.0 mm,100. mm,125. mm,150.
mm,300. mm,500. mm,1000 mm,2000 mm,4000 mm, 8000mm,
Sphericity,Blocks,Intensity
Or
Year,Month,Day,Hour,Minute,Second,D10,D20,D25,D30,D50,D70,D75,D80,D90,D95,D
99,Xc,Nros,Xmax,#200,#140,#100, #80, #60, #50, #40, #30, #20, #16, #10, #8, #4, .25
in, .375 in, .5 in, .75 in, 1.0 in, 1.5 in, 2.0 in, 2.5 in, 3.0 in, 4.0 in, 8.0 in, 16 in, 32 in, 64
in, 128 in, 256 in, 512 in, Sphericity, Blocks, Intensity
56

WipFrag User's Manual PDF

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

WipFrag User's Manual PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

WipFrag User's Manual PDF

Uploaded by

Copyright:

Available Formats

WipFrag 2.

WipFrag© and WipJoint© are the property of WipWare Inc.

WipWare Customer Support:

WipWare Technical Support:

Questions concerning software performance should be addressed to:

4.3.4 Block Delete – F5 16

APPENDIX 4: DEFINITIONS OF WIPFRAG STATISTICS 41

APPENDIX 5: FILE NAMING AND GRAPH LABELING CONVENTIONS 44

APPENDIX 6: WIPFRAG SECURITY KEY 45

APPENDIX 7: EXAMPLES OF WIPFRAG OUTPUT 46

APPENDIX 8: IMAGING MOVING PICTURES 47

APPENDIX 10: OPTIONAL WIPJOINT MODULE 49

APPENDIX 13: WDM DRIVERS/V. 2.4 INSTALLATION 54

1.1 Fragmentation Measurement

Blast fragmentation measurement is a recent addition to an extensive family of

1.2 The WipFrag Method

WipFrag is designed for two main types of input:

2. SETTING UP THE WIPFRAG COMPUTER

2.1.1 Desktop Workstation Installation

2.1.3 Other Preparations

Power up the computer and install and start WipFrag.

3.1 WipFrag Window

3.2 WipFrag Operations

3.3.1 Basic Operation

1. Capturing a new image or open a disk image.

2. Set the scale on the image.

3. Activate the automatic edge detection algorithms.

4. (Optionally) Use the editing tools to improve the edge

5. Do a virtual sieve, to measure block sizes and generate a graph.

3.3.2 Basic Setup

4. Options / Metric/Imperial. Selects units of measurement, meters or inches. This

4. WIPFRAG ACTION MENUS

4.1 Main Menus and Toolbars

4.1.1 Opening Menu

4.1.2 Image View Menu

4.1.3 Image View Tool Bar

Only the first, second,

4.2 File Menu and Tools

4.2.2 Open Multiple

4.2.3 Capture New Image

4.2.3.1 Capture new image dialog

Preview - Displays a “live” image.

4.2.7 Open Net

4.2.8 Save Net

4.2.10 Print Preview

4.2.11 Print Setup

4.2.12 (Recently used file list)

4.3 Edit Menu and Tools

1. WipFrag allows manual editing to insert missing boundaries between fused

4.3.2 Draw Line – F2

4.3.3 Draw Polyline – F3

4.3.4 Block Delete – F5

4.3.5 Delete Line – F4

4.3.7 Ignore Block – F6

4.3.8 Mark as Fines – F7

4.4 View Menu and Tools

4.4.1 Edit Net

Gaussian Standard Deviation is the strength of the Gaussian smoothing convolution.

Provide uniform indirect or diffuse lighting without excessive sharp or one-sided

For underground photography, choose area illumination with several medium-power