As an important component of ITS, Video Surveillance System is widely applied in traffic environment, especially in urban transportation in recent years. The existing video monitoring systems encounter two difficulties. First, complex traffic scenes cause dense vehicles and various vehicle driving road conditions. Second, with deeper study of ITS, video monitoring system available today is not content with basic offerings, but rather more complex applications. The main contributions of this paper include: 1. Firstly, this paper proposes a novel hierarchical and modular surveillance system (HMSS). The fundamental starting point of this paper is that we believe that the real traffic world is a complex system, so the fundamental problem of video surveillance system is how to recognize and understand the complex system visually. 2. Secondly, this paper presents a deformable template---meaning Active Basis Model (mABM), which is utilized to detect vehicles in road segments. As every element uniquely corresponds some specific part contour of a car, the mABM not only detects vehicles but also recognizes vehicle contours. 3. Thirdly, the compositional deformable model is presented to detect vehicles in crossings. The motivation of CDM is that a vehicle object can be decomposed into components such as roof, bonnet, windscreen, etc. The CDM can effectively detect the vehicles with different types, contours, sizes and poses. 4. Fourthly, the proposed method of vehicle detection in this paper is applied to the actual system. The core technique of our video surveillance system is that vehicles can be located in various traffic scenes. The system detects vehicles by combining the features of vehicle contour, license plate and vehicle lamp, and achieves two functions. The one is traffic information collection system, the other one is electronic policing system.
修改评论