Research On Moving Vehicle Detection And Tracking Algorithm Based On Video Processing

In recent years,with the rapid development of image processing technology,artificial intelligence and pattern recognition and other disciplines,computer vision has also received a lot of attention,and video technology has been widely used in various fields.Urban traffic problems caused by the rapid increase in the number of cars have become more and more serious.Therefore,in the field of intelligent transportation,moving vehicle detection and tracking technology based on video processing has become a hot research topic,and it is of great significance to carry out research on moving vehicle detection and tracking algorithms.This article focuses on the research of moving vehicle detection and tracking algorithms in video surveillance.The specific research content is as follows:(1)Several commonly used detection methods for moving vehicles are introduced and analyzed and compared,and an improved vehicle detection method is proposed by fully combining the advantages of each detection method.This method firstly combines the five-frame difference method with the Prewitt edge detection operator to remove excess noise;secondly,for the problems in the mean background modeling process,a dynamic background update method is proposed,and a mask is innovatively designed,Use part of the current image to update the background image,reducing the amount of calculation while reducing the sensitivity to weather changes;On this basis,an improved vehicle detection method combining five-frame difference and average background models is proposed.Comparing experiments under five different road traffic scenarios,the results show that the moving vehicles detected by the method in this paper are more complete and accurate,can adapt to different scenarios,make up for the shortcomings of the inter-frame difference method and the background difference method,and have better real-time performance.(2)In the aspect of moving vehicle tracking,a moving vehicle tracking algorithm based on multi-feature fusion is proposed.The traditional KCF algorithm only extracts HOG features and cannot be applied to complex scenes.On this basis,this article combines HOG features and CN color features to perform feature extraction to achieve complementary advantages and add scale change estimation strategies to make up for it.The KCF algorithm tracks the disadvantage of a fixed frame size.Through the tracking and comparison experiments of moving vehicles in two different road traffic video scenes,the experimental results show that the method in this paper can accurately track moving vehicles,and the method is feasible and effective.(3)Judge the vehicle retrograde behavior in the vehicle violation phenomenon.According to the specific location of the tracked vehicle,the center of mass coordinates and movement track information of the vehicle can be obtained.Through the movement trajectory of different vehicles in the same video and the change of the longitudinal coordinate of the vehicle’s center of mass,it can be judged whether the moving vehicle has retrograde behavior.