Rockwell Automation 5370-CVIM2 Module - Manual (Page 194)

5 Chapter Chapter 6
Reference Tools
6–35
The First Pass panel contains several selection fields, data entry fields, and
buttons, which are described briefly, as follows:
•
Masking –– This field selects Enabled or Disabled for the masking
function. The default is Disabled .
•
Stop When –– This field selects either the “ Best ” or the “ First ” match of
the feature in the search window to the stored feature image in the feature
window. The default is “
Best .”
•
X Scale –– This field selects the scaling ratio for pixels along the X–axis
in the feature and search windows. The default is 1:4.
•
Y Scale –– This field selects the scaling ratio for pixels along the Y–axis
in the feature and search windows. The default is 1:4.
•
Scale To –– This field selects either the “ Nearest Neighbor ” or the
“ Neighborhood Average ” as the basis for the scaling function. The
default is “ Neighborhood Average .”
•
Ignore
Pixel Errors –– This field is used to adjust the sensitivity of the
reference window tool. It determines how closely the template must
match the image on a per pixel basis. The default value is 10.
•
Max. RMS
Pixel Errors –– This field determines the average
acceptable deviation between the template and image. The default value is
64.
•
Define Mask –– The button activates the Threshold/Filter
function, which is used to “define” the mask on the feature image.
•
Done –– The button saves the currently selected configuration
settings for the First Pass function, then exits the First Pass panel and
returns to the tool edit panel.
Masking Function
The masking function can be used to “mask” edge pixels within the feature
image or “template” when those pixels are not only unnecessary for
template–matching purposes, but create errors large enough to cause the
reference window tool to fail.
In general, masking is likely to be needed whenever a feature image contains
many edge pixels relative to the number of pixels in the feature image (area).
These edge pixels may yield large matching errors when the image is
compared to a corresponding image in the search window, and the reference
window tool may fail because of the matching errors along the edges.
Figure 6.30 (page 6–36) provides two examples to illustrate when masking is
needed (A) and when it may not be needed (B).
In example (A), the feature image contains many sharply defined edges. The
edge pixels in this image can cause large matching errors, and the tool may
fail. With masking enabled, however, the edges of the feature image can be
ignored, and the likelihood of a match increased accordingly.