Fast Moving Targets Segmentation of Traffic Video Based on Between-Class Variance Method
基于最大类间方差法的快速交通视频运动目标分割方法

Moving object segmention in traffic video requires a large amount of information and a long computational time. A new fast moving object Segmentation method is proposed. Two appropriate images are sclected. Then, the absolute value of subtraction is taken. This paper proceeds to block images and decides which block images should be given up. The remaining sub-blocks for otsu are chosen, and the moving objects are segmented. Experimental result shows that this algorithm can greatly improve the speed of moving object segmentation.