Research On Moving Objects Detection Algorithms In Complex Scene

Moving objects detection is a very active subject in computer vision, which has very wide application, especially in the intelligent video surveillance system where the first step is moving objects detection.This paper from the practical set out actually, research on background model build and moving objects detection based on the video image sequence.We first analyze the present typical background difference algorithms, and then introduce some different background models using in background difference, including Gaussian mixture model and Non-parametric method etc. We analyze the advantages and disadvantages of the algorithm and their scenario. According to background difference method will not be applied in sudden changed illumination scene, and can't effectively reduce and eliminate the interference of shadow, we put forward and introduce a background difference method based on the light effect statistical model. Through the experiments show that the algorithm can effectively deal with the influence of light mutations, and can improve the robustness of the background difference method.In addition, since the present background difference method is not suitable for the dynamic scene caused by slight camera shake, we put forward a moving objects detection algorithm based on Gaussian mixture model and saliency map, through studying the basic principle and model of the visual significant attention mechanism, which will not be affected by shake but have the interference of the stationary object significant, Experiments show that this method can effectively avoid the interference caused by slight camera shake and eliminate the influence of the shadow.