Research On Assistant Navigation For The Safe Steering Of Vehicles

Vehicle technique is a pop field of ITS. The research on road recognition and obstacle detecting, which is a key technique for intelligent vehicle navigation system, does great contribution to safe driving and even unmanned driving. A monocular-based guiding system for in-front vehicle detecting, tracking and overtaking is designed according to real condition of traveling. An algorithm for detecting and tracking roads and barriers in self adapting way in various backgrounds is proposed to guide the overtaking.A method based on monocular vision is proposed to detect the lane marking in foreground to acquire the relative position and direction of the lane and in-front vehicle. The information is analyzed to prompt the driver. In the thesis, the algorithm is analyzed, studied, improved and simulated.The filtering algorithms are analyzed and a proper one is improved to remove the noises and enhance contrast. On the basis of pre-processing, the level demarcation line between the road surface and the background is found out by the projection method, and then the region of interest (ROI) is defined. The feature of the lane is extracted by searching and tracking the edge spots. The straight road model is built according to people's driving experience, prior knowledge and mathematic description. Then the characteristic points of the lane marking are matched to the corresponding model by least-squares fit. The in-front cars in different foregrounds, white or black, are detected by filtering in the frequency plane and enhanced by morphologic method. The algorithm is compensated by reducing the ROI window. The estimate of distance of the in-front car is analyzed by combining the information of vision and radar sensor. Then the analyzed video is supplied. The algorithm is adaptable for detecting different vehicles in different weather. A circulating method is designed in the thesis to dealing with a large number of videos in different backgrounds. The efficiency is highly advanced by reducing searching window of current image according to the result of the previous frame.The system is adaptable in the circumstance of both-lane structured road with vehicle as the barrier. Most of the algorithms in the thesis have been validated by the videos captured from real traffic road scenes, and the experimental results show that the methods are efficient, stable and accurate.