Systems and methods for classifying and aligning highly similar or self-similar patterns

摘要

A method of training an image processing system to create a geometric model, the method comprising the steps of obtaining a first set of training images of a first workpiece and a second set of training images of a second workpiece; selecting at least one training image of one of the first set of training images and training at least one seed level matching model from the at least one seed image; detecting the training images in the first set of training images that were not selected as the at least one seed image using the at least one seed value; An adaptation model for obtaining a first set of relative positions to the at least one output image for each training image in the first set of training images; acquiring the training images in the second set of training images not having selected the at least one seed image using the at least one seed matching model to obtain a second set of relative positions to the at least one seed image for each training image in the second set of training images; identifying first corresponding features from the first one A set of training images, the first corresponding features being stable among the training images in the first set; identifying second corresponding features from the second set of training images, the second corresponding features being stable among the training images in the second set; identifying at least one common feature among the first corresponding features and the second corresponding features; extracting one or more distinguishing features from the first set of training images and the second set of training images based on the first corresponding ones Features, the second corresponding features, and the at least one common feature among the first corresponding features and the second corresponding features, the one or more distinguishing features between two sets of training images corresponding to features between the first corresponding features and the second corresponding features are different and can be used to differentiate between the first workpiece and the second workpiece; creating an alignment model using at least one of the first corresponding features, the second corresponding features, and the at least one common feature; andcreating a classification model using the one or more distinguishing features.