Research On Target Tracking System Based On Vibe And Optical Flow Method

Target tracking technology is one of the key technologies in intelligent video surveillance,providing information such as the position and moving trajectory of target objects in the video.Target tracking is the premise of follow-up monitoring applications such as target recognition,and has gradually become one of the research hotspots in the field of computer vision.However,video surveillance has disadvantages such as complex background and frequent occlusion,which leads to a series of difficult problems.This paper mainly studies the background modeling and detection efficiency of surveillance video.The main research contents are as follows:1.An optimized optical flow algorithm based on model matching and motion prediction is proposed.Optical flow method is a classic target tracking algorithm.It does not require prior conditions,nor does it need to model the background and moving targets.The tracking results contain the target's motion information.However,this algorithm is inefficient and difficult to process video in real time.Aiming at the above problems,an optimized optical flow algorithm is proposed:using the motion information and motion prediction of the moving target to reduce the impact of the occlusion problem;and the tracking points of the same target are divided into the main body and other parts,and the matching model reduces the number of queries The number of tracking points is consistent with the target motion information query,which improves the efficiency and accuracy.2.An optimization algorithm based on background modeling is proposed.Vibe algorithm is an efficient and accurate image analysis algorithm for extracting non-background objects.However,since the sample set of all pixels is stored in this process,the sample set of new pixel values and historical values of the samples for each subsequent frame,and the sample set is updated to determine whether it belongs to the background point,so a static or slowly moving object Being absorbed into the background,and moving objects will be affected after leaving the background collection.In order to reduce the impact of this problem,starting from the update strategy and suppressing the update of the moving target neighborhood,an optimized vibe algorithm is proposed,and the algorithm simulation and analysis are performed.3.A multifunctional surveillance camera is designed that can be widely used in streets,schools,companies and other safe neighborhoods.The camera is energy-saving and environmentally friendly,and is easy to install and use.It can pre-process the captured video,eliminate jitter,reduce noise,and facilitate subsequent algorithms to identify and track targets in the surveillance video.Figure[33]Table[3]References[63]...