Research on Intelligent Transportation Vehicle Detection and Tracking Algorithms Based on Video

By means of sequence mean method based on a series of consecutive video images to extract the background, this paper makes use of difference model of colored images to make difference on both images, hence to obtain effective goals of moving vehicles. As regard to post-treatment of the test results, it proposes an algorithm for real-time segmentation and integration of the moving target. The real-time segmentation mainly adopts an improved algorithm based on the elimination of the shadow, to divide the moving target and by matching the template, separate the overlapping vehicles. In the time of integration, this essay brings forth a type of interest function to make a real-time filling of the internal hollowness of the moving target, and have a realtime merger by making use of the distances between the rectangles. Experimental results have shown that the various proposed algorithm is characterized by simplicity, easy-to-use and high real-time. The vehicle detection rate is over 96%.