Research On Wheel Alignment Technology Based On Vision Measurement

With the rapid development of auto industry, vehicle ownership is increasing rapidly. Automobile has become the essential means of transport in our work and life. But with the increase in automobile mileage and the use of the extension, the state of technology is bound to change. The operating stability and traffic security, abnormal tire wearing and increasing fuel consumption of the vehicle are related to inaccuracy of the wheel alignment. At present, most wheel alignment equipments use contact-type method at home and abroad. Because of limitation of measurement principle, the inspection process is complex. So it can't realize rapid inspection.Computer Vision is the physical structure of extracting and studying three-dimensional scene through the two-dimensional images of perception for three-dimensional world. As a new and advanced high-precision measurement technology, Computer vision measurement technology has been in the traffic navigation, field measurement, production automation, virtual reality and many other areas of a successful application.In response to these circumstances, computer vision measurement technology is used to the wheel alignment in this paper. The measurement of wheel alignment parameters based on computer vision technology is elaborated, and a new system solution is designed to measure wheel alignment parameters. Tessellated template is used for camera calibration reference in this paper. Corners of template are extracted by the measurement of Harris corner detection. Control points coordinates can be achieved by the homography between templates and images. Through camera calibration based on tessellated template, intrinsic parameters and extrinsic parameters and homography matrix are achieved. Homography is adopted to acquire the world coordinates of the target from image coordinates, and then determine the spatial location of the target, which is used to acquire necessary angle.Experiments are carried out to verify the wheel alignment system model based on computer vision measurement technology. The results show that the model is feasible and correct, and there are broad prospects for its application to the automobile wheel alignment detection.