Research On Expressway Traffic Incident Detection System Based On YOLO-v3

With the gradual improvement of China’s highway network,there are more and more traffic incidents on highways,and the importance of highway traffic safety is highlighted.The timely discovery,handling,and rescue after traffic incidents have also become important issues for highway managers.The detection of traffic incidents on highways has always been a key research direction for scholars studying traffic safety.In recent years,with the popularization of highway surveillance cameras on roads,the detection of traffic incidents through image recognition has also become a research hotspot.This article combs the detection and processing flow of traffic incidents on Lianxu Expressway,builds an auxiliary detection system for traffic incidents on the basis of the existing hardware and software of Lianxu Expressway,and focuses on the research and development of the data processing module of the system.To detect and alarm traffic incidents.The thesis first studies the highway incident detection system,and then proposes to establish a data processing module based on image processing technology.The core algorithm of the data processing module is the target detection algorithm and target tracking algorithm.The first is to use the traditional background difference method target detection algorithm to detect the moving vehicles in the video image,and then use the improved Camshift tracking algorithm to track the moving vehicles,so as to detect the running state of the vehicle in the video and timely abnormal traffic conditions deal with.The traditional algorithm has realized the detection and tracking of moving vehicles,but it is unable to cope with the video of heavy traffic and shaking.The thesis also studies the characteristics of the moving target of the highway,and selects the data processing technology based on the YOLO-v3 algorithm to realize the detection and tracking of the moving target,and can realize the alarm function in the event of traffic jams and traffic accidents.In the data processing module based on YOLO-v3,this paper uses the Deepsort algorithm for pedestrian tracking,and is trained in the VOC2007 data set to complete the detection,tracking and alarm of the vehicle.The research data of this paper is collected from the real data of the monitoring center of the G30 Lianxu Expressway Section.The experimental algorithm uses software such as Python as the programming platform.The research results show that the highway traffic incident detection system studied in this paper can complete the detection of traffic incidents in the video To improve inspection efficiency and provide guidance for highway safety management.