Research On Motion Target Detection And Tracking Algorithm Based On Video Sequence

Video target tracking plays an important role in the field of computer vision,and has broad application prospects in many fields,such as intelligent transportation,public safety,.artificial intelligence and so on.However,there are many problems with the traditional target tracking algorithm,for example,the impact of the environment is relatively large,when the moving target occlusion prone to track the loss can not be re capture the target.How to track the moving target effectively and accurately has always been a concern in the field of computer vision.In this paper,the moving object detection algorithm based on codebook algorithm is studied at first,and the principle and performance characteristics of the algorithm are given next.In order to solve the problem that the original codebook algorithm has slow computation speed in moving object detection,this paper proposes a Codebook moving object detection algorithm based on color space improvement and parameter optimization.The original codebook algorithm is converted from RGB space to YUV space.The three color space channel was reduced to one after analysis,while using the brightness difference instead of the original maximum and minimum brightness parameters,by introducing the codeword weight coefficient to delete and optimize other parameters,so that the codebook algorithm is optimized to guarantee high accuracy at the same time greatly improves the computational speed of the algorithm in moving object detection.In order to solve the problem that the original TLD algorithm is prone to tracking drift in the tracking process,this paper proposes a TLD algorithm based on key feature points,which is called STLD.It contains abundant information of the feature points to replace the Grid uniform sampling in the original TLD algorithm.The tracking accuracy of the moving target is improved,suppression of the original TLD tracking the drift,but it can also reduce the sampling points of the track loss rate.So it has better inhibitory effect of drift and faster speed,improve the robustness of the algorithm.In order to solve the problem of tracking failure caused by occlusion or deformation of the original TLD algorithm,this paper proposes a STLD algorithm based on Kalman filter which is called KSTLD.The predictor is introduced in the front of the STLD algorithm detector,and the target position is predicted by the kalman predictor to enhance the correlation of the moving target position in the frame before and after the video sequence.We can obtain the approximate region of the position of the target in the current image.The predictor results are combined with the three-stage classifier in the STLD algorithm detector to improve the detection effect of the moving target in the occlusion environment and improve the accuracy and operation speed of the STLD algorithm detector.