Fast image matching algorithm based on template partitioning and subtemplate matching

Abstract

Matching algorithms based on pixel grey value are very popular and widely used in image matching. However, these algorithms have high time complexity and are sensitive to the size of image. To reduce the time complexity and sensitivity to the size of image of the matching algorithms based on pixel grey value, a fast image matching algorithm based on template partitioning and subtemplate matching is proposed. The algorithm divides the template into multiple subtemplates, then the eigenvalue of every subtemplate is calculated and the optimal one is chosen when the difference between the eigenvalues of the subtemplates and the subimage is minimum, and finally the best matching position for the reference image is determined by adopting non-ergodic search during image matching processing. The algorithm's time complexity is reduced to O((S − N + 1)² × N²). Experimental results show that the proposed algorithm is superior to the circular projection algorithm and the NCC algorithm and has low computational complexity.