| As the accelerating of China’s urbanization, cars have become one of the main transportation facilities in our daily life. But the traditional transport system can’t meet the needs of traffic management because of the increasing of car ownership. As a result, the negative effects become increasingly prominent, such as traffic congestion and frequent accidents, etc. Currently it is necessary to find an effective way to reinforce traffic management to avoid the traffic accidents and to reduce the mortality. Since the intelligent transportation system based on the image processing has a feature of being timely and efficient, it has attracted more and more attention.This paper mainly researches the algorithm of peccancy monitoring in intelligent transportation system. Two methods recognizing whether the driver is wearing safety belt and making a phone call during driving is designed and the detection process of peccancy vehicles is improved. Because all those recognition methods are detected in the vehicle window area, the accuracy of the window location has a great impact on the subsequent detections. So this paper focuses on the window location analysis and design at first, and then researches an algorithm about window location based on edge detection and integral projection. In order to improve the accuracy of detecting whether the driver is wearing safety belt, a method of seatbelt detection that based on Hough transform is redesigned through the research of the Hough transform algorithm and the analysis of lines’ various conditions detected with Hough transform. In view of the existing skin color models are targeted at all kinds of images for segmenting skin color, a kind of skin color segmentation model utilizing the face detection is researched based on the actual situation. Firstly, Viola-Jones face detection algorithm is used to detect the human face, and the skin color model is then constructed by using the pixel points in the face region as samples. And a method based on Viola-Jones algorithm is designed to detect if the driver is calling during driving. In order to improve the whole monitoring system efficiency, a way of peccancy vehicles detection is improved. In process of constructing the lane line model, straight line is used for lane lines’ linear approximation. During the process of tracking moving vehicles, the additional vehicles in the lane are matched with the vehicles that are disappearing in the adjacent lane of the previous frame only when they are not driving along the sign direction.Finally by considering user’s requirements and system performance, a complete set of peccancy monitoring system that mainly includes the driver violation detection and vehicle peccancy detection is designed. The system can verify that the recognition method of the driver if he/she is using the safety belt or making a call during driving has a high detection rate and processing efficiency. |
