About The Research To Information Identification And Processing System Of Obstacle For Unmanned Vehicle

The intelligent vehicle to achieve unmanned, automatic and safe driving as the ultimate goal is a more rapid development research field in recent years.The following work,taking information identification and processing system of obstacle for unmanned vehicle as object of study, object,focuses the research work on path detection,obstacle recognition and the system implementation.In the aspect of the hardware circuit design of system,power modules, sensor modules and the DSP peripheral interface circuit of the whole system are designed.In the aspect of software design of system,the road detection and obstacles identification algorithm are studied.The main coments and contributions of this thesis are as follows:(1)The overall program of this thesis is designed, the working principle of system and the function of main part are analyzed.(2)It is designed that power , DSP chip peripheral circuits, video sensors, wheel speed sensors, angle sensors, acceleration sensors, ultrasonic sensors, temperature sensors of the overall system, providing support for software design and implementation.(3)The methods of road detection and obstacle recognition based on video processing are discussed. They are composite analyzed that the current variety of image preprocessing, road detection and obstacle recognition technology. A number of practical methods are proposed,such as: the noise caused by external environment interference are eliminated or attenuated making use of median filter; the white lane is extracted from the background through the binary segmentation algorithm of higher accuracy; the sampling points of lane mark make use of Hough transform to identify; according to the angle scope of lane mark and the distance between left and right lane, the sampling points are filtered; sample points are fitted a straight lane track line with linear interpolation and the least square method; the bottom edge of the obstacle which detected by pixel gray level is considered as the bottom shadow of vehicle; obstacle recognition is achieved using of texture features.(4)The algorithm for image preprocessing, road detectionand,obstacle recognition, are carried out software design and programming.Finally, the simulation of software algorithm was introduced in MATLAB 7.4 platform. The results show that proposed road detection and obstacle recognition technology have better robustness and reliability, It can extract the exact path information and the obstacle feature point.