| Ensuring the traffic run smoothly, detecting the traffic incidents automatically and dealing with them are vital. In this paper, we will dissertate on the topic of detecting the traffic incidents automatically based on digital imaging processing technology. And a software based on the theory is also given out.The main content and contribution of this dissertation are summarized as follows:1. Traffic incidents happened on the expressway are categorized and especially the category of traffic accidents is elaborated in detail. The main target of this detecting theory is the accidents taken place on the expressway and the traffic incidents caused by the accidents.2. The demands of the foreground detecting and the scene of the expressway are considered. Firstly, picking up of the foreground is based on the method of background difference. Secondly, choosing a proper median filter is significant to suppress the noise interference. Finally, the method of gray threshold, which is ground on the matrix histogram, otsu guidelines and the curve-fitting of analytical expression method, is helpful to the extract of the foreground object on the expressway.3. The attribute of connected region demarcated by the binary image makes get the characteristics of the foreground objects in the traffic incidents easily. The vehicles on the expressway are classified as car, bus, van. Vehicles-Dimensions of these groups are further subdivided. The area threshold and aspect ratio threshold in the given detecting area are referenced in the judgment of the traffic incidents.4. The velocity model and proposal which is based on the video image captured on the expressway are developed. What the model needed is the changes of the coordinates and the time difference between two appointed frames. In this way, detecting the velocity of the vehicles and the traffic incidents such as traffic jam and drive slow are more accurate. Furthermore, this model can also report incidents caused by the traffic accidents. |
PDF Full Text Request