The Research On Detection And Tracking Algorithm Of Moving Target In Video Image Sequence

Computer vision is widely used in application software,digital equipment and industrial robots,which realizes the work that can only be completed by human eyes.Moving target detection and tracking is the core part of computer vision algorithm.Although domestic and foreign scholars have done a lot of analysis and research on two algorithms,the environment of the moving target located in the actual scene is often complex and changeable.Various interference factors will affect the accuracy of target detection and target tracking.The existing algorithms can not be applied to all kinds of complex and changeable scenes,therefore the shortcomings of the target detection and tracking algorithms are improved.(1)In view of the traditional moving object detection ViBe algorithm,it is slow to eliminate ghost shadow,easy to form drag shadow,easy to miss detection and easy to be interfered by noise.When the state of motion of an object changes slowly,it will leave a shadow and other problems.An improved ViBe algorithm combined with the three frame difference algorithm is used to eliminate ghosts by improving the initialization mode of the background model and the update rate of the background.The adaptive threshold is used to improve the detection effect.The noise and hole in the background are removed by morphological processing.(2)When the Camshift algorithm is used for target tracking,it can not eliminate the interference of objects with the same color feature in the scene and the moving target,and the failure of target tracking after the target is blocked.Therefore,this paper adopts the Camshift algorithm that integrates AKAZE feature matching and Kalman filtering to improve the Camshift algorithm.When there are some objects interfering with the moving target in the scene,it will lead to the failure of tracking the moving target.Adding the AKAZE feature matching algorithm can accurately locate the moving target.When the target is blocked,the Kalman algorithm can predict the position of the target and update its parameters in time to achieve accurate tracking of the target.(3)The design and implementation of video monitoring software mainly adopts VS2015,Qt5.9 and Opencv3.1 as development tools.The improved algorithm is applied to the system to realize target detection and tracking.In order to enhance the function of the system,a part of the laboratory project is added to the system.