The Intelligent Car Control System Based On Visual Feedback

Smart car, also known as a wheeled robot, has environmental awareness,information feedback, remote control and other functions together and can work in theenvironments which humans can not adapt to, so it has been widely applied in variousfields of industry, military, families, hospitals and so on.This paper selected a smart phone with Android system to achieve effective controlof the four wheel car through wireless communication. The small car is equipped withmonocular vision unit, which composed by camera, frame grabbers, and imageprocessing unit. Frame grabbers can capture the image of the traffic information aroundthe four wheel car through the camera, and then transmit the information to the imageprocessing unit, at last sent the picture to the smart phone by the WIFI wireless module.The smart phone sends commands that based on the car’s real-time traffic information tocontrol the car’s running.This article describes the overall framework of the smart car system which basedon monocular visual feedback and complish the functions of the various components ofthe system. For the problem of the images collected in the car movement being not clear,we deeply study the point spread function of uniform linear motion blur of theimage,then compare a variety of image restoration algorithms and select a frequencydomain method that can identify the key parameters (the blur direction and blur length)of the point spread function. This paper completes the process of algorithm on Androidplatform, to make the feedback image more clearly and confirm the safe running ofsmall car.