Hardware manual

Extreme Difference Impact Reference Guide
Datalogic Automation Inc. 3-150
value. Also, since location is unimportant, non-maximal suppression is not used. Defect points tend to clus-
ter and larger clusters indicate larger defects.
The Extreme Difference tool uses concentric square regions to compare regional extremum (minimums and
maximums). For example, if the tool is searching for dark regions, the minimum grey levels for the interior
square and the exterior ring are calculated. These two values are compared and marked as a defect if the
exterior minimum value minus the interior minimum value exceeds the difference threshold. Extremum val-
ues are used, rather than average or median values, so that the tool is insensitive to edges.
The interior region must be large enough to contain any defect that is to be detected. This means that the
defect must be completely contained within the inner radius. If the defect is larger than the inner radius, the
defect will not be found. This is because, when the defect extends into the outer ring it raises or lowers the
outer ring’s minimum value, affecting the difference calculation. Also, the tool works more effectively when
the defect sizes are consistent.
For example, in this picture the red dots correspond to pixel values used in the interior region and the green
dots to the exterior region. The Inner Radius is three and the Outer Ring Width is two.
Input Name What it is
Input Image Linked Image
Tool Origin The tool’s origin relative to the entire image space
Find Shape List This is the tool’s search area.
Shape List ROI
Type
Each Shape is separate ROI: Each shape is processed as a separate
ROI and the results are combined. This mode is faster than the other
mode if the shapes are far apart, but results are duplicated if the shapes
overlap.
ShapeList is One ROI: The ShapeList is one large rectangular ROI and
the shapes are the enabled pixels within the rectangle. The tool may be
slow if the shapes are far apart. This setting allows shapes to overlap
and form complex shapes without repeating the overlapped pixels and
duplicating results.
Inner Radius The inner square’s radius. The inner square edge size is (2 times the
radius) - 1. For example, if the radius is 3, the inner square edge size is
5. The defect must be completely contained in the inner radius in order
to be detected. It can not extend into the outer ring.
Outer Ring Width The width, in pixels, of the outer ring.The outer square edge size is twice
the outer ring width plus the inner square edge size.
Difference Thresh-
old
The minimum allowed difference between the inner and outer extreme
grey levels (in percent). For dark objects, the difference is the outer min-
imum minus the inner minimum. For bright objects, it is the inner maxi-
mum minus the outer maximum.
Interior (red dot) square
Inner Radius = 3
Exterior (green dot) ring
Outer Ring Width = 2