Detailed Explanation of AOI Algorithm - Image Statistical Modeling Algorithm
The image statistical modeling algorithm is a dedicated detection algorithm for ALeader and is applied in almost all detection fields. AOI statistical modeling enhances the system's ability to recognize OK and NG images by learning a series of OK templates, observing image changes, and combining the visual biases seen in all OK images to identify the features of component shape changes and possible future change patterns. Its algorithm flag in the detection algorithm is "OTHER". In the process of learning the OK template, the following three problems are mainly solved:
What should the shape of component A be like?
That is, the size, shape, color and surface pattern of the components, etc.
What changes will occur to Component B?
That is, the variation rules of the natural dimensions, shapes, colors and surface patterns of the components.
How much will the shape of component C change?
That is, to what extent the changes in the size, shape, color, surface pattern, etc. of the components are reasonable.
Finally, what is obtained is a standard model for testing that integrates the above elements and lies between OK and NG.