Vision-based Research On Aided Navigation Of Intelligent Vehicle

Aided navigation technique is one of the key techniques of intelligent vehicle, while the technique of pedestrian detection and tracking is currently one of the most active research topics in the domain of aided navigation of intelligent vehicle. This strong interest is driven by a wide spectrum of promising application in many areas such as virtual reality, smart surveillance system, advanced user interface, motion analysis and model-based coding, especially aided navigation of intelligent vehicle. Research on it does great contribution to the development of intelligent vehicle techniques and safety of city roads.This dissertation is focused on the pedestrian detection and tracking in one-eyed video images. Firstly, the methods of the motion detection of pedestrian in video images are analyzed. Edge of differences algorithm to detect the pedestrian is proposed. The algorithm combines space segmentation and motion segmentation. Secondly, the motion tracking of pedestrian in video images is studied. Based on the boundary model of body, the boundary tracking algorithm is used to track the pedestrian in video images. Finally, a system of pedestrian detection and tracking is realized. In the experiments, the monocular vision grayscale video sequences are obtained by a hand-held video camera used to replace the video camera equipped on the moving vehicle. The designed system consists of two function modules: the pedestrian detection module and the pedestrian tracking module. Much of the algorithm proposed and used has been based on analysis of pre-recorded videos of real traffic road scenes and proved to be effective and efficient. At the end of this paper, some viewpoints and suggestions about future work are given.