Vehicle Detection Based On Video

With the rapid development of society and economy, the number of vehicles has quickly increased. The transportation hugs, traffic accident and environment pollution increasingly threaten our environment and economy development. Under this background, the intelligence transportation system (ITS) establishes, and has acceptedly become the best path of solving transportation problem. In ITS, vehicle detection based on video is a pivotal technique that becomes research focus.Firstly, this paper briefly introduces development background and actuality of ITS and summarizes actual research difficulty. Secondly, in the foundation of analyzing deficiency, we mainly study vehicles segmentation, shadow elimination and parameter measure and so on. Finally, we write code, imitate experiment and analyze result. Concretely speaking, as follows:(1) Moving vehicles segmentationThe segmentation of image with infirm contrast is very difficult. We propose a new fast threshold segmentation arithmetic based on the technique of the minimum interclass variance and region growing. We firstly make difference, and then improve the method of minimum interclass variance, which is used to obtain the optimal segmentation threshold of difference image. Finally, the target is segmented by the region growing method.(2) Shadow of vehicles eliminationShadow of vehicles can influence the elimination effect and induce falsely detection. On the basis of shadow feature, we propose a method for shadow elimination based on edge extraction and chroma information. Experiments indicate that the method is feasible.(3) Traffic parameter measureFinally, we thoroughly study overspeed peccancy action in traffic and propose material overspeed determinant-detection arithmetic.