An Image Feature Matching Algorithm with Clustering Constraints

Aiming at the problem that the traditional ORB feature matching algorithm has many mismatches and poor robustness, a feature matching optimization algorithm with clustering constraints is proposed. Firstly, the K-means++ algorithm is used to assign classification labels to the feature points obtained in the current image, and the clustering region is redivided. Secondly, the corresponding clustering region is found in the next picture, and the feature points in the region are also assigned classification labels. Then, the classification labels of feature points in the two images were compared during matching, and the inconsistent matching point pairs were removed. Finally, the RANSAC algorithm is used to optimize the matching results, and the results are verified. The results show that the proposed algorithm can effectively eliminate the false matching points, and the running speed is greatly improved.