Research On Multiple Moving Objects Detection Algorithm Based On The Video

In computer vision and intelligent video surveillance community, the higher levels such as moving object classification, tracking and behavior understanding heavily depend on the results of moving objects detection. Moving object detection is one of the fundamental tasks for video analysis, which active topic widely is researched in the last decade around the world. In this paper, we focus on the detection of moving objects in complex environment, and place emphasis on some background models and shadow detection especially.This paper first summarizes the domestic and international status quo of the moving objects detection technology and the shadow detection technology. And the various methods of moving objects detection and moving objects segmentation are researched. In the analysis of the advantages and disadvantages of the traditional moving objects detection methods, the background difference method is used to detect moving objects.Four background modeling algorithms are widely researched and analyzed based on extensive experiments. The comparison conclusions are presented. An improved background model update algorithm is presented to stabilize the choice of the parameters. And the dynamic thresh is get by the method which using histogram. Experiments show that the algorithm can detect moving objects more accurately and deal with some gusty circumstances.Finally, the main algorithms of shadow detection are systematically researched. Furthermore, the ones in gray space and color space are classified and compared each other through experimental analysis. In color space, the algorithm in RGB and HSV respectively detect shadow through analyzing the shadow characteristics in each color space. Experiments show that the shadow detection algorithm in RGB space has a higher shadow detection rate, but the shadow recognition rate was lower than the algorithm in HSV space. In gray space, with the absence of color information, the algorithm based on Gabor dictionary can extract the shadow texture features. It can detect shadow by using the similar shadow features between the background point and the shadow point. Experiment results show that the algorithm has good shadow detection rate and shadow recognition rate.