Vehicle Detection Method Based On Binocular Vision And Image Process Technology

This paper describes an on-road objects detection method which is based on a binocular vision system. At first, for increasing the image process speed, canny algorithms and Hough transform are used to recognize road area. Then, vehicle can be distinguished by shadow extension. At last, the distance relations of each object in the image will be obtained. Not as some traditional methods which are using radar and infrared ray, the ultimate aim of our thesis is to detect the forward road's objects' condition and distance, and though this we can decrease the cost and improve the method. We operate this research using hardware such as two cameras, and software such as MATLAB and OpenCV. The extensions are contributing the results to the traffic jam and driving safety research. The result shows that all main objects and distance relations of them can be detected accurately.