Advertisement
Smart Matching
9.1
Smart Matching
Deviation processing
As shown in the diagram below, deviation processing is used when you only want
abnormalities outside the object workpiece to be recognized as differences, but not
differences in the outline area of the object workpiece.
Template
The Imagechecker will use the deviation data for each pixel in the results obtained
during the subtraction processing, and will revise the portions you do not want it to
recognize as differences (the outline area in the example above) to the results below
the threshold value. By doing this, the corresponding portion will appear as the final
results of the binarization. Deviation data is set for each pixel in the template.
As shown in the diagram below, with subtraction processing alone, differences in the
outline areas (difference: 128 > threshold value: 96) are recognized as black parts
in the second row, but if deviation data is also used, the results of the outline portions
become equal to the threshold value (difference: 128 × deviation:
0.75 = 96 threshold value), and are not recognized as abnormalities. On the other
hand, abnormalities that exceed the threshold value (difference: 128 × deviation: 1.0
= 128 > threshold value) are recognized as such.
Area where the outline
fluctuates due to the
workpiece. Difference = 128
9 − 12
Want this to be judged OK.
Want these to be judged NG.
Deviation = 0.75
Deviation data
Deviation = 1.0
OCR Type A230
Advertisement
