Computer vision technology based Intelligent Video Surveillance (IVS) is widely used in our daily life. Pedestrian counting, as one of the most important applications in IVS, is of great significance in mall security, business data analysis and so on. In this thesis, we study the pedestrian counting methods in the static camera scenes, and design a real-time pedestrian counting software system.First of all, we introduce several related theories of computer vision, and analyze the existed pedestrian detection and tracking algorithms. Then, we propose a pedestrian detection method which is a combination of the HOG detection method and the background subtraction detection algorithm. Furthermore, we study the pedestrian tracking algorithm based on HOG feature. Finally, in accordance with the above method, we design and implement a prototype of video pedestrian counting systems. Experimental results show that our method can achieve the accuracy and real-time demand for pedestrian counting application. |