空间光学技术研究室知识库

In this paper, a full information vector recognition algorithm for moving targets is proposed on the basis of the characteristic distribution of point targets and the moving characteristic between frames. The traditional multi-frame image fusion method of moving target recognition is abandoned. We utilize the distribution characteristic of point targets extracted from single image and moving characteristic of point targets extracted from multiple images to recognize and classify moving targets with the similarity principle of feature vector. Compared with the traditional maximum likelihood estimation image processing algorithm, the proposed recognition method costs less computation and provides a novel approach for spatial moving target detection and recognition. ? 2019 SPIE.