Research On The Key Technologies Of Real-Time Visual Analysis Of Moving Object

Moving object detection and tracking is currently one of the most active research topics in the domain of computer vision. It aims at attempting to detect, track and identify moving objects, and more generally, to understand their behaviors. It has a wide spectrum of promising applications in many areas such as virtual reality, smart surveillance, perceptual user interface, etc.Computer intelligent video surveillance system(CIVSS) is one of new arising high-tech application fields. It spans many subjects including computer science, machine vision, image engineering, pattern analysis, artificial intelligence, etc. CIVSS can automatically analyse sequence image by the methods of computer vision and video analysis. The system can real-time detect, recognize, and track moving objects in a special environment. Furthermore, it can also analyse and judge the behavior of objects. The aims of CIVSS are to understand the meanings of video stream and to explain the scenes comprehensibly, hence to guide the action and make some decision.The research of this paper mainly applies in the field of CIVSS, and this topic comes from international cooperation projects. This paper mainly discusses the detection,segmentation,tracking and counting of moving human in video sequences gathered from a fixed camera. First, this paper uses two effectual method of moving object detection: one is combining background subtract with connected components labeling, and another is combining temporal difference with bi-directional projection histogram. The advantages and disadvantages of the two algorithms is analysed and compared. Shadow may appear under different scene and time, moreover shadow will influence the veracity of object detection,classification and tracking. This paper implements several algorithms of shadow removal, and the advantages and disadvantages of these algorithms is analysed and compared, finally this paper uses the standard rgb model of RGB space,the luminance model of YUV space and the chroma and saturation model of HSV space to restrain the influence of shadow preferably. This paper proposes an algorithm to dispose occlusion among objects called bi-directional projection histogram, and this method is effectual. This paper analyses the geometrical and pixel distributing feature of each object region to implement object classification(people and vehicle). Finally, an algorithm based on the value function is developed to track and count object. This paper designs an real-time and effective system for counting the crowd, and the precision of this system is above 85%, furthermore this system is applied in Japan.