Design Of Automatic Vehicle Obstacle Recognition System Based On ARM

With the rapid development of science and technology, the study on obstacle avoidance of embedded intelligent car has attracted more and more attention gradually.In order to improve the smart car driving safely in unknown environment, the machine vision technology has been applied to the automatic obstacle recognition system in this article. Extracting road information which determines an obstacle in front of the car promotes development of the smart car technology.In this article, based on the ARM(Advanced RISC Machines) platform, the recognition system of obstacles of smart car has been made a deeply exploration and research. The machine vision technology was used to change the collected road environment information into image signal, and the automatic obstacle recognition technology was combined with it effectively. And then through software simulation and practicality experiment on smart car in unknown environment to avoid the obstacle was verified. Through the research of automatic recognition system of obstacles, it is an aim to provide security for the future development and application of the smart car technology.According to the specific requirements of the practically application of machine vision in the smart car, the various parts of the obstacle recognition system has been designed. ARM S3C2440 is adopted as the control processor of the obstacle recognition system, and the hardware platform framework including the module of the main control board and core board, the USB camera image acquisition module, the DC motor driver module and the power supply module of the system, etc is built. Then, on the basis of the hardware platform, the software platform of the automatic vehicle obstacle recognition system based on ARM platform is built, and the embedded cross compile environment is established successfully, and the transplantation of theOpen CV and Qt graphical interface library is completed. Then the road environment image information is obtained by the image acquisition module, and the target obstacles are extracted by region growing, so as to indentify the road conditions and the location relationship between the obstacles and the smart car. And the obstacle avoidance algorithm is obtained according to the possible several location relationship between the smart car and the obstacles by the thought of region segmentation. And next, the control program of the machine vision processing and the autonomous obstacle avoidance under the Linux system is written. Finally, the design of the vehicle obstacle automatic identification system is applied in this paper, in a real environment with the experiment platform. And the experiment of the ARM smart car avoid obstacles is completed automatically. It has proved that the proposed algorithm of obstacle avoidance in the actual is effective, and some problems which appear in the process of obstacle recognition are analysed and the corresponding countermeasures to solve these problems are proposed according to the experimental results.