Research On Video Surveillance Technology Of Moving Vehicles Based On Android Mobile Terminal

With the development of modern intelligence,the Intelligent Transportation System(ITS)has an important role in solving traffic problems,real-time vehicle detection and tracking is one of the research hotspots in ITS.Vehicle detection and tracking can be performed all the time by installing inspection equipment in the past,however,it is not possible to detect vehicles at any location due to the fixed nature of the equipment,so the research of moving vehicles of video surveillance based on android mobile terminal is proposed by this article,which plays a great significance for studying the collection of traffic information in ITS.The main research contents are as follows:(1)Foreground target detection based on improved Vibe background model algorithm.The foreground target detection of the moving vehicles is the presupposition of achieving vehicle tracking,the subject analyzes and compares the advantages and disadvantages of the target detection methods such as background subtraction method,optical flow method and temporal difference,the background difference based on Vibe background modeling can not only accurately extract the moving foreground targets,but also has a faster processing data speed,which is suitable for the field of moving vehicle detection.The traditional Vibe algorithm has the disadvantage that the ghost region can not be eliminated for a long time when moving vehicles are detected,so this paper proposes a novel improved Vibe algorithm that fuses the mean background frame.Experimental results show that the improved Vibe algorithm can eliminate ghosts more quickly.(2)The research of Kalman filter tracking algorithm based on feature matching.Considering that the vehicle detection system is eventually applied to the mobile platform that has low performance relative to the computer,it is not appropriate to use very sophisticated tracking algorithm,finally,a Kalman filter tracking algorithm based on feature matching is adopted.The Kalman filter is used to predict the positions of the foreground targets' centroid of the previous frame in the next frame,these predicted positions are matched with the new frame's foreground objects,the successfully matched targets in the previous frame indicate that the targets tracking is successful,and the targets that did not match successfully are added to the vehicle count to complete the vehicle counting function.(3)The implement of cross-platform transplantation technologies,the vehicle detection program implemented on the PC using the O technology.Through in-depth learning of the Android system platform and other related penCV open source computer vision library is transplanted to the Android platform,which not only increases the flexibility of Vehicle detection,but also expands the application area of the Android platform.