Template Matching Algorithm Based on Edge Detection

A new template matching algorithm is proposed to improve rotation invariance of mean absolute difference method. The matching function is composed of two sub functions. The first sub-function matches the gray difference between template image and object image, and the second sub-function matches the edge intensity difference between them. The template image and object image are arranged into circle structure, which ensure not only the rotation invariance of the algorithm but also detailed information considered. Simulation results show that the method can realize the target recognition function, and the recognition accuracy can be raised effectively.