System and method for image pattern matching using a unified signal transform

A method for performing pattern matching to locate an instance of one or more of a set of template images in a target image, wherein each template image comprises a plurality of pixels, the method comprising:

• determining a unified signal transform from the set of template images;
• applying the unified signal transform for at least one generalized frequency to each of the set of template images to calculate a corresponding at least one generalized frequency component value for each of the set of template images;
• receiving the target image;
• performing pattern matching using portions of one or more of the template images and the target image to determine possible template image locations in the target image;
• applying the unified signal transform for at least one generalized frequency to one or more portions of the target image at the possible template image locations in the target image to calculate a corresponding at least one generalized frequency component value for the possible template image locations in the target image;
• determining a best match between the at least one generalized frequency component value of the portions of the target image and the at least one generalized frequency component value of each of the template images;
• and
• generating information indicating a best match template image from the set of template images for each of the portions of the target image.

A computer medium comprising program instructions which are executable to locate an instance of one or more of a set of template images in a target image, wherein each template image comprises a plurality of pixels, wherein the program instructions are executable to perform:

• determining a signal transform for the set of template images, wherein the signal transform is operable to convert an image into a form comprising generalized basis functions, wherein the generalized basis functions represent an algebraic structure of the set of template images, and wherein at least two of the template images are uncorrelated;
• calculating one or more values of the signal transform applied to each of the set of template images at at least one generalized frequency;
• receiving the target image;
• calculating one or more values of the signal transform applied to the target image at the at least one generalized frequency;
• determining a closest match between the one or more values of the transformation of the target image and the one or more values of the transformation for each of the set of template images;
• and
• generating information indicating a closest match template image of the set of template images.

A machine vision system, comprising:

• an image acquisition device for acquiring a target image;
• a computer system coupled to the image acquisition device, wherein the computer system includes a processor and a memory medium, wherein the memory medium stores a software program executable to locate an instance of one or more of a set of template images in the target image, wherein each template image comprises a plurality of pixels, wherein the software program is executable by the processor to perform:
• in a preprocessing phase:
• determining a unified signal transform for the set of template images;
• calculating one or more values of the signal transform applied to each of the set of template images at at least one generalized frequency;
• and
• in a runtime phase:
• receiving the target image;
• calculating one or more values of the unified signal transform applied to the target image at the at least one generalized frequency;
• determining a closest match between the one or more values of the transformation of the target image and the one or more values of the transformation for each of the set of template images;
• and
• generating information indicating a closest match template image of the set of template images.