Moving targets detection algorithm based on the temporal-spatial segmentation algorithm

Abstract

To detect moving targets in an image sequence fast and accurately, a temporal-spatial segmentation algorithm of moving objects is addressed in this paper. First of all, the motion compensation is used in the received low frequency sub-images of a consecutive sequence that three frames images which is transformed with the 2D wavelet. After that, the temporal segmentation algorithm is used to extract the areas of the moving targets in the sub-images, and then the accurate areas of the moving targets in the original images are extracted with the spatial segmentation algorithm. To raise the efficiency of the algorithm, the Bisection method is introduced into the spatial segmentation method. It is shown by the experimental results that the moving targets are detected fast and accurately with the algorithm in the paper, and that the occlusion problem is solved well in the multi-target motion.