System or method for detecting defect within a semi-opaque enclosure

A method for detecting an object of unknown shape, configuration, and location on an inner wall surface of a rigid semi-opaque enclosure having a semi-opaque first side portion having first internal and external surfaces and a semi-opaque second side portion having second internal and external surfaces, said first side portion being different from and opposing said second side portion, said first external surface having a pattern visible externally to said enclosure, said method comprising:

• directing first channel light toward and onto said first external surface of said semi-opaque enclosure;
• capturing, from portions of said first channel light reflected off said first external surface but not having passed through said first internal surface, a first digital image composing data representing said pattern;
• directing second channel light toward and onto said obiect and toward and onto said first side portion of said enclosure;
• capturing, from portions of said second channel light emanating from said first external surface after first traveling along side and around said object inside said enclosure and then passing through said first internal surface, a second digital image comprising a silhouette of said object;
• and
• forming a difference image representing said object and substantially devoid of information representing said visible pattern by subtracting one of said first and second digital images from the other;
• wherein said rigid semi-opaque enclosure comprises a bottle and a semi-opaque label attached to and covering a substantial portion of an outer side-wall surface of said bottle, said substantial portion of said outer side-wall surface including said first external surface;
• wherein said first and second digital images comprise two-dimensional gray level images represented with respective two-dimensional light intensity function f.sub.1 (x, y) and f.sub.2 (x, y), wherein x and y denote spatial coordinates and f.sub.1 and f.sub.2 are values proportional to the gray level of the respective image at a point x, y;
• wherein said different image comprises a two-dimensional gray level image, said method further comprising performing a resegmentation operation on said difference image, said resegmentation operation comprising forming a two-dimensional binary image from said difference image;
• wherein said forming of a two-dimensional binary image from said difference image comprises performing a thresholding operation on said difference image, said thresholding operation comprising forming a two-dimensional binary thresholded image g (x, y) in accordance with the following:
• ##EQU1## where x and y are spatial coordinates of a point (x, y) of said difference image, d (x, y) is the tray level of said point (x, y), and p (x, y) is a local threshold control variable corresponding to said point (x, y);
• wherein p (x, y) represents a gray level local threshold image defined as a function of at least one of said first and second digital images f.sub.1 (x, y) and f.sub.2 (x, y);
• and
• wherein p(x, y) is obtained by performing a local threshold mapping operation on said second digital image f.sub.2 (x, y), said local threshold mapping operation comprising performing point processing in the spatial domain on said second digital image f.sub.2 (x, y) in accordance with
• r=dmap(s), for a pixel having spatial coordinates (x, y)
• where dmap is a mapping function, r is a gray level equal to p(x, y) and s is a gray level equal to f.sub.2 (x, y).