Font Size: a A A

Moving Object Detection And Tracking In Video Images

Posted on:2018-05-31Degree:MasterType:Thesis
Country:ChinaCandidate:F PanFull Text:PDF
GTID:2348330542961775Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
The research of moving object detection and tracking technology in video images is a hot spot in the field of computer vision.At present,the research on this subject has been widely applied to many fields such as security monitoring,intelligent transportation,human-computer interaction,and medical diagnosis.In this thesis,the moving object detection and tracking methods in video images are studied.The main research content is as follows:In the aspect of moving target detection,a moving object detection algorithm based on improved Harris corner and OPTICS clustering is proposed.Firstly,the Harris corner detection operator with scale information is added to extract the feature points of the adjacent images.Then,the feature points of the adjacent images are matched to get the motion vector of the feature points,and the optical flow graph is established according to the motion vector.Finally,the improved OPTICS algorithm is used to complete the clustering on the optical flow graph,and the moving target is accurately separated from the background of the image.In the aspect of moving target tracking,Mean Shift moving target tracking algorithm is studied,and an improved algorithm which improves the tracking effect on scale change,rotation and occluded moving target is proposed.In the improved algorithm,the Kalman filter is used to predict the candidate position of the moving target,then takes the candidate position as the initial position for the Mean Shift iteration to obtain the new candidate position,and finally the Bhattacharyya coefficient is used to judge whether there is occlusion in the image.If the occlusion occurs,the candidate position predicted by Kalman filter is treated as the tracking result.If there is no occlusion,the candidate position of Mean Shift algorithm is taken as the position of the moving target,and then uses the change matrix of the feature points to change the scale of the moving target area and obtain the final motion target tracking result.At last,the moving target detection and tracking system based on ARM is developed,the key technologies of the system's hardware and software is described,and the real-time monitoring,target detection and target tracking function of this system are tested.
Keywords/Search Tags:Image processing, Moving target detection, Target tracking, ARM
PDF Full Text Request
Related items