Research On Detection Method Of Vehicle Illegal Behavior Based On Dome Camera

Recently, with the rapid increase of vehicles, traffic congestion and traffic accidents become very urgent problems. The traffic violation will seriously affect the normal operation of the urban traffic system. We study the detection method of traffic violations based on the dome camera, and design a traffic violation detection and snapshot system.Firstly, we study the problem of moving object detection in traffic surveillance, and introduce a background modeling method based on visual background extractor. Meanwhile, we use a vehicle detection method based on HOG features and SVM algorithm to distinguish between vehicles and non-motor vehicles. Secondly, we study the methods of moving vehicle tracking. For the specific application, we propose a region-based feature matching and tracking algorithm. Furthermore, we apply this algorithm to detect vehicle illegal converse and illegal parking. The detection of traffic violation can be achieved accurately through our approach. Finally, on the basis of the proposed algorithms, we design a traffic violation detection and snapshot system based on the dome camera. The system performance is evaluated in a real scene. The result shows that this system achieves good effects in detection efficiency, capture accuracy and system stability, which is an effective solution for vehicle illegal converse and vehicle illegal parking problems.