Research On Motion Vehicle Detection And Tracking Method Based On Improved Frame Difference Method And Camshift Algorithm

Intelligent video surveillance is used more and more in the work and life of people at all levels of society,its convenience that gives for people is self-evident.The detection and tracking of moving vehicles is the basis technology of the intelligent detection and monitoring system in the intelligent transportation system.The purpose of moving vehicle detection is to extract the interest region in the video sequence,so as to extract the corresponding target features for the subsequent tracking.The purpose of the moving vehicle tracking is to know the movement characteristics of the vehicle in real time.The work of vehicle detection and tracking has laid a solid foundation for traffic incident detection and behavior analysis.Due to the complex scene of target detection and tracking,the unpredictable factors,angle and distance of the camera,greatly increased the difficulty of target tracking.Therefore,it is of great significance to study the methods of moving object detection and tracking.In this paper,the following research work is carried out on the detection and tracking of moving vehicles on highways.(1)A moving object detection method based on five frame difference and LOG operator is proposed.First of all,the two frame difference,the difference of the three frames,the difference of the four frames and the difference of the five frames are studied and simulated,the experimental results show that the five frame difference detection gets the best result.Then,combining the LOG edge detection operator that can detect motion target complete information and the difference of the five frames together are used to detecting the moving object.The experimental results show that the new algorithms have good adaptability to the environment,the effect of drying is good,and the information of target detection is complete.(2)Moving target tracking method based on regularized particle filter to improve Camshift.Camshift algorithm is mainly based on the target color distribution to trak,so it can achieve the tracking goal with a relatively large difference with the surrounding environment.But it is easy to have a similar color background or environmental interference resulting in the loss of tracking.Therefore,an improved Camshift algorithm is proposed based on the regularized particle filter method,the method can realize the good processing for the new target and the similar target combining the moving target detection method,and it realizes the lasting and stable tracking.This paper based on frame difference and Camshift study the method of moving vehicle detection and tracking taking the expressway as the research object.It enrich the research contents of foreground object detection and tracking in surveillance video,and it can be used as a reference for foreground object detection and tracking in other scenes.