Adept AdeptSight User Guide - Manual (Page 294)

Histogram
classification of valid blobs.
The Blob Analyzer returns an array of numerical results for each valid blob that has been
found and located. Blob results include geometric, topological, and grey-level properties.
Histogram
For every new image, the Blob Analyzer generates a histogram representing the distribution
of the pixel values in the region of interest. Pixel values in the histogram range from 0 (black)
to 255 (white).
Thresholding
Thresholds are used to segment the image into two types of pixels: blob pixels and back-
ground pixels. Depending on the segmentation mode selected, either one or two thresholds
can be used.
There are two types of thresholding functions:
Hard Thresholding
Hard Thresholding (also referred to as binary thresholding), segregates pixels into one of two
states: 0 for background pixels, and 1 for blob pixels. The result is a binary image (for exam-
ple, a white blob on a black background). Hard thresholding assumes that changes in data
values occur at the boundary between pixels, without allowing for a variation in grey-level
values across blob boundaries. Because this is rarely the case, soft thresholding (described in
the next section) is more often used for applications.
Soft Thresholding
Soft Thresholding provides flexibility in treating pixels that border blob regions. A soft thresh-
old is sloped and covers a range of pixel values. Processed pixels within the threshold range
are output as weighted pixels. Weighted pixels are used to calculate blob results in a pro-
portion corresponding to their value in the threshold range.
l The range of values within a soft threshold is user-defined and corresponds to the dif-
ference between a maximum and a minimum threshold value.
l All weighted pixels appear in the image. However, they contribute to the property
results in proportion to their weight in the threshold range.
l The Pixel Weight Image displays the weighted pixels in corresponding levels of grey.
Dynamic Thresholding
Dynamic Thresholding provides an adaptive thresholding mode, which is useful when there
are lighting variations from one instance to another. A dynamic threshold sets a percentage
of the pixel distribution (histogram) to be considered as blob pixels.
AdeptSight User's Guide, Version 3.2.x, Updated: 8/23/2012
Page 294