The Research On Moving Object Detection And Tracking In Complex Scene

With the rapid development of computer technology,intelligent video surveillance technology has been widely used in people's daily life,and has brought great convenience to our life.The main problem of the intelligent video surveillance technology is the detection and tracking of moving targets.It's an important topic in the field of computer vision.It has been widely used in many fields,such as artificial intelligence,pattern recognition,image processing,medical imaging and so on.It is an important part of many visual applications,such as intelligent video surveillance,human-computer interaction,intelligent driving,etc..The changes of the target itself and the background environment are challenge in the complex scene.Therefore,the research of moving object detection and tracking in complex scene has very important study value and practical significance.(1)As the Gaussian mixture model algorithm uses a fixed update rate,and can't detect the moving targets accurately in complex scenes.An improved algorithm based on Gaussian mixture model is proposed,the algorithm combined the improved frame difference method with Gaussian mixture model.In this new algorithm the background is divided into different regions,and different regions use update rate,the background model can adapt to external interference in complex scenes better.The experimental results show that the improved algorithm proposed in this paper can adapt to the interference of the external environment such as illumination change and tree occlusion in the complex scene,and improve the accuracy of the detection.(2)As the particle filter algorithm abandon the particle with low weight in the process of resampling,an improved particle filter algorithm based on firefly algorithm is proposed.In this algorithm,we use the firefly algorithm to optimize the particle distribution in the process of importance sampling.In the process of resampling,the firefly algorithm is used to simulate the firefly moving to the high likelihood region,and update the global optimal value which can improve the effectiveness and diversity of particle.The experimental results show that under the influence of many factors,such as the change of background,the irregular movement and the occlusion of trees,the improved algorithm can use less particle and track more efficiently and accurately.(3)In this paper,we design and implement a video surveillance system based on MFC interface library and computer vision library in VS2010 environment.The system has the function of video input,image preprocessing,moving target detection,morphological processing,target tracking and trajectory drawing.The target detection and target tracking module use the proposed detection and tracking methods.