Detection And 3D Trajectory Tracking Of Points Feature Group Motion

With the development of computer technology and image processing technology in recent years,the detection and tracking of moving targets has been gradually applied to intelligent transportation,human-computer interaction,medical diagnosis and treatment,industrial control and other aspects.In order to obtain the position information of the object,at the same time,predict and recover the target's trajectory,the better way to do that is to detect,analyse and process the tracked target in video image by vision-based target detection and the technology of tracking computer.This paper discusses the problem of group motion detection and trajectory tracking with point feature.The current factors affecting trajectory tracking are target detection,matching and tracking.The article analyzes and summarizes the target detection and tracking technologies existing at this stage.Based on this,one matching method based on Euclidean distance and one tracking method based on color feature are proposed.The large-scale optical flow algorithm is optimized and improved.This paper proposes a method for detecting and trajectory tracking of point feature group motion.These main researches are as follows:(1)The theoretical background and related technologies involved in the research are introduced in detail.It includes an overview of binocular stereo vision systems,optical flow techniques,several common feature detection and matching algorithms,classical tracking methods,and theoretical foundations for 3D reconstruction.This part provides theoretical support for experimental research.(2)The purpose of stereo correction,technical classification,and implementation steps are specifically described.The target center is detected by ellipse fitting,and a feature matching method for point feature group is proposed based on Euclidean distance.The spatial reconstruction is performed using triangulation after the feature matching is completed.(3)Analyze and summarize traditional target tracking methods,such as Meanshift algorithm and Camshift algorithm.The optimization and improvement of the large-scale optical flow tracking algorithm improves the tracking accuracy of the point feature group and completes the trajectory reconstruction on the time series.Finally,the motion data such as displacement and velocity are calculated and analyzed,and the physical characteristics of the target motion are analyzed.The experimental results show that the method proposed in this paper is better than the SIFT and SURF algorithm,and the matching time is shorter.Compared with these traditional target tracking methods,the improved large-scale optical flow tracking algorithm can effectively improve the accuracy of reconstruction.The above experimental results prove the effectiveness and feasibility of the proposed method.