Rockwell Automation 5370-CVIM2 Module - Manual (Page 206)

5 Chapter Chapter 6
Reference Tools
6–47
Note that the Second Pass panel is a scaled down version of the First Pass
panel. It has only the Masking and Stop When selection fields and the
“ Ignore ” and “ Max. RMS ” pixel error parameter entry fields. X and Y
scaling is fixed at the 1:1 ratio. The same defaults are used. For information
about these parameters, refer to the First Pass Configuration section.
When Double Pass is enabled (default), the second pass operates on the
results created by the first pass operation. The principal advantage of using
two passes is the achievement of high resolution in coordinate and angle
results without the significant speed penalty that usually occurs when using
one pass with high resolution scaling (1:1) for X and Y axis scaling.
During the second pass, the feature window searches only in the part of the
search window in which the first pass located a feature –– it does not search
the whole search window –– thereby saving considerable time in many cases.
Table 6.2 demonstrates the speed difference for one feature (A or B) using
the same feature image and search image for single pass and double pass.
These results are derived from an actual setup in which the “pick and place”
coordinates of the feature image were 120 and 143 for the X–axis and
Y–axis, respectively.
Table 6.2 Example: Comparison of Single Pass and Double Pass Performance
Operation Scaling Masking
Coordinates
(X, Y)
Time for
“Best”
Time for
“First”
Single Pass 1:1 for X and Y axes Enabled 120, 143 1980ms 350ms
Double Pass:
First Pass
1:8 for X and Y axes Enabled 119, 139 Not used 10ms
Double Pass:
Second Pass
1:1 for X and Y axes
(not user selectable)
Enabled 120, 143 28ms Not used
Note that when using single pass with high resolution scaling (1:1 for X and
Y axes), the coordinates were located at 120 and 143, the original locations,
but the processing times were quite long: 350ms for “First,” and 1980ms
(almost two seconds) for “Best.”
However, when using a double pass with low resolution scaling (1:8 for X–
and Y–axes), the first pass requires only 10ms to locate the feature
approximately, and the second pass requires only an additional 28ms to
locate the feature more precisely. Note that in this example the coordinates of
the located feature are the same as in the original feature image: 120 and
143. The total time is 38ms for the double pass operation vs at least 350ms
for the single pass operation.
NOTE: If you are using a reference window for rotation compensation,
return to page 6–30 and repeat the configuration process for Feature B.