Research On The Method Of The Front Obstacle Recognition Based On Binocular And Its Application

With the improvement of China’s national comprehensive strength and people’s living standard, the automobile has become one of the main transports, maintaining a rapid growth rate. During enjoying the convenience by automobiles, people are suffering numerous questions such as environment deterioration and many traffic accidents which caused great harm to people’s life and threaten the security of our property. In order to reduce the occurrence of traffic accidents effectively, the recognition system for the road forward the car using auxiliary driving has become a hot research direction all over the world. The key point of this system study is the method to improve the accuracy of identifying the location of the vehicle, measure the distance to the obstacle by location algorithm and remind the driver of barrier in time about to crash to avoid accident.Aim of the main reason to pilot error of traffic accident, the project proposed a method to recognize the obstacle based on binocular on the road forward the car. The theory of binocular vision and artificial intelligence, and the technology of image processing can be used to identify the position of vehicles on the road ahead accurately. After measuring the distance using binocular algorithm of obstacle, system will remind the driver in time about to crash to avoid accident if the measured distance is smaller than the safety distance.This paper studies the structural road environment like highway using binocular stereo vision system. Using video processing technology to obtain the lane line and region of interest, and then pre-judge suspected obstacles in the region. After the discovery of suspected obstacles, rejudge the obstacles. If confirm the position of the obstacle after rejudgement, measure the real-time ROI obstacle distance using reconstructed binocular 3D coordinates to provide real-time traffic information to the driver. Based on the minimum safety distance threshold algorithm, remind the driver to evade the operation timely when there is a danger of accident to avoid the accident occurrence. Store up the image information by large capacity hard disk, for the query or evidence obtaining afterwards. Upload the information to the PC through the wireless GPRS, for real-time analysis and processing using. This paper uses the actual picture taken to describe the algorithm of vehicle system; uses matlab to carry on the simulation experiment and finally prepare and optimize the C language algorithm and transplant it to the DSP hardware platform. The experimental results show that the measurement precision of the ranging system can satisfy the safety distance measurement needs of in road traffic, which is a kind of effective recognition system for detecting the road status before moving vehicles.