Improved K-means Template Matching Target Recognition

Target recognition mainly includes feature extraction and target classification based on features. This paper mainly improves the traditional target classification algorithm based on K-means template matching to improve the target recognition rate. In view of the problems that the matching template established by the traditional target classification algorithm based on K-means template matching cannot accurately represent the distribution of recognition targets, the clustering center is fuzzy, and the recognition rate of some targets is low, this paper proposes an improved target classification algorithm based on K-means template matching. In the process of template establishment, more subclasses are added, and the mapping between subclasses and recognition targets is established. Thus, the problem of fuzzy clustering center is solved, and the target recognition rate is improved. Compared with the target classification algorithm based on SVM, it can achieve better target recognition rate with lower computation.