Research On Optical Flow And Its Application On Video Motion Detection

sequences. Motion detection in video sequences is an important field of computer vision research topic. Video is made in time series of images related to the composition, and its wealth of raw data, the strong correlation of adjacent frames and time varying mode of the dynamic characteristics of the image of the moving target is easy detection, segmentation and recognition. Video motion detection, not only in industrial and military applications with a very important, but also in a number of important public places, such as ports ,square, calaboose , road vehicle monitoring system has a wide range of applications.The process of motion detection based on optical flow includes image preprocessing, optical flow computation, the utilizing optical flow computation, motion object extraction.The main works are as follows:1. In the actual environment, the light and the object and other reasons, often the optical flow calculation by the shadow of the interference, optical flow is detected object is not accurate, this paper presents a gradient of image brightness difference to remove the shadow. Allows more accurate detection of objects.2. The traditional optical flow based on differential calculation on the image brightness is not conserved, and the movement is not continuous, the accuracy of optical flow calculation will be greatly decreased, which are contrary to the traditional optical flow assumptions. Black & Anandan proposed to solve the brightness is not conserved and discrete issues such as movements BA algorithm. This article will go to the shadow algorithm and the combination of BA algorithm, a method ---- GSOF enhanced optical flow algorithm.3. GSOF algorithm will be applied to road vehicles were detected and the number of statistics, progress had proved the effectiveness of the algorithm.The paper is organized as following five parts; Section 1 gives the introduction of the background, significance and development of the optical flow computation. Section 2 introduces some classical methods in optical flow computation. Section 3 introduces the BA optical flow computation and a algorithm of the moving cast shadow detect and remove. The Section 4 using the GSOF algorithm to detect the road vehicle. The Section5 applies to the GSOF algorithm in people count. In the last section introduces the application to the abnormal human behavior identification.