The Method of Image Matching by Taking Every Fixed Match Pixel

Abstract

A method of image matching by taking every fixed match pixel was proposed, then the images could be matched accurately in case of noise. the method was based on gray feature-based template matching, it took a long time for basic template matching to traverse to extract match points. According to this problem, the method that took every fixed matching pixel in template image and real-time image, and took every fixed pixel to search in real-time image was used to achieve image matching. the results show that: for noise-free image, under the premise of guaranteeing the matching accuracy, this method can achieve the same effect as the Sequential Similarity Detection Algorithm does in matching speed, for noise image, the speed of image matching is improved by this method, and the problem that Sequential Similarity Detection Algorithm can't guarantee the matching accuracy is solved at the same time.