Moving Target Detection And Tracking Research Based On OpenCV

Moving object detection and continuously tracking technology is now a hot research direction and the leading edge areas of computer vision, because of its integration of Computer Science, Control Theory, Machine Vision, Image Process, Pattern Recognition, Advanced Mathematics and other subjects of advanced technologies.And it is widely used in the intelligent monitoring, industrial inspection, the medical analysis, and military and many other industries, because it lies behind the enormous economic and business opportunities, it has attracted lots of scientific institutions, academia and business attention and many domestic and international universities, scientific institutions, companies to invest heavily in the field efforts research and exploration, and they have made a large number of achievements. The topic of this paper is to make deeper exploration and experimentation on the basis of these results. Because of this, focus of this paper is to explore the theory of object detection and object tracking.Tracking and detection of moving objects use computer vision to analyze image sequence, implementation of dynamic scene moving object positioning, identification and tracking, and be on the basis of identity, obtained moving object video sequence understanding in order to guide various practical applications. Because of its disadvantages, the traditional method of moving objects detection is difficult to satisfy different environments and the speed of simultaneously detection and accuracy requirements, especially for complex light environment, at the same time, moving object detection and tracking results are not very ideal. Just for this situation, a bit of work to do in this article to resolve this problem.(1) In the part of object detection research, Focus on the background subtraction method and continuous difference method and integrate the two typical algorithms together and do an adaptive background updating algorithm that can adapt to changes in light and environment backgrounds. The image is detected using mathematical line morphological, the method of segmentation results are processed to eliminate noise caused by interference.(2) Pretreatment of the image have summary, on this basis, this paper proposed a method of experimental environment suitable image enhancement, image noise processing and image segmentation, to improve the classic maximum class pitch algorithm, moving object detection tracking is based on the principle of Kalman filtering which is derived deeply, this article has given a solution, and then the experimental results were horizontal and vertical analyzed and compared.