Method Research For Moving Object Detection With Complex Background

In this paper, based on the background of video processing technology and moving objects detection technology, with the comparative analysis of three commonly used objects detection algorithms (frame difference, background subtraction, optical flow), the adavantages, disadvantages and applicability of algorithms are summarized. According to complex scenes like illumination change, background disturbance, the abnormal motion of foreground or background objects, target slow motion, foreground color similar to background color and so on, the new detection algorithms corresponding to above scenes are proposed or improved. And these new algorithms'validity is tested and proved through algorithm demonstration platform.In this paper the significance, research status and hot issues of moving objects detection with the complex background are summarized and the basic principles and applicability of three commonly used objects detection algorithms are discussed in detail at first. Then several complex backgrounds are concluded and combined and brought into our research area. At last, based on the research analysis of existing solutions about the given complex background, the new detection algorithms are proposed or improved, at the same time, the algorithms'applicability and limitations are analyzed and discussed.In the last part of this paper, an algorithm demonstration platform is made up, which uses VC++ and OpenCV. We can verify the validity of the new algorithm and get the research's future development direction by this demo result on this platform. This moving object detection method research with complex background has some guidance and reference for the future real-time video analysis.