Research On Moving Target Detection And Target Tracking Under Occlusion

The increasing demand for security in the information age has led to the rapid spread of video surveillance and has promoted the rapid development of video surveillance in various fields such as transportation,security,and military.Moving target detection and tracking technology as an important research content of video surveillance,the improvement and breakthrough of this technology is of great significance to the development and application of video surveillance.For the problems that occur in the application of video surveillance,the thesis research on the technology of moving target detection and tracking,the effectiveness and accuracy of the target detection process are improved by the proposed algorithm,and the tracking of moving objects in the occlusion environment is realized.The related research of this thesis mainly includes the following two aspects:In the research of moving target detection,the background subtraction method is susceptible to interference such as light and results in a low target detection accuracy.The accuracy of the background model is related to the effect of moving target detection.Create an initial background based on the average background,update the background model based on background accumulation,and combine it with the three-frame difference method,take experiment on moving target under multiple video sequences.Experimental results show that the recognition rate of the moving target detection algorithm reaches 97.3%.In the research of moving target tracking,for the failure of the MeanShift algorithm when the target is occluded during the tracking process,Kalman filter which can predict the state of the moving object is proposed to improve the MeanShift algorithm.The simulation results show that the improved algorithm can effectively track the occluded target in the video.