Detection System, Based On Computer Vision Vehicle Wheel Alignment

This thesis studies the four-wheel positioning detection of the automobiles by using computer vision technology. Through the processing of the target board images captured by CCD camera, the related data and parameters can be obtained. The four-wheel positioning detection, based on computer vision technology, can detect automobile wheel positioning parameters in a simple, rapid and accurate way, and has a good prospect of application.First, the thesis studies the current structures and measurement principles of four-wheel positioning devices, analyzes the measurement principles of currently used e-cable positioning device and light-beam positioning device, and points out their shortcomings. Then it uses relevant geometric theories and perspective theories to analyze the measuring principles of wheel positioning parameters which are based on computer vision. Through detailed analysis of the methods based on perspective and space vector, the thesis gives the mathematical model of wheel positioning parameters, improves the target board design based on perspective mode, and puts forward improvement measures for traditional detection based on perspective mode,that is, the distance measurement is completed by the computer vision system, which can simplify the existing structures, improve the detection efficiency and reduce costs. This thesis also calibrates the video cameras for computer image acquisition and processing system, and obtains their internal and external parameters. Finally, this thesis makes experimental verification and error analysis to test the mathematical model and its algorithm. The results show that the proposed four-wheel-detection method is right. And the thesis analyzes the errors in the experiment and puts forward some measures to reduce those errors.Aiming at computer vision technology in the course of automobile wheel positioning performance testing, this thesis mainly focuses on solving the following questions:1. Calibration of the camera in the four-wheel positioning detection by using computer vision technology.2. The principles and algorithm of vector test and perspective test in the four-wheel positioning detection by using computer vision technology.3. The improvement of target board used in the four-wheel positioning detection.this thesis makes comprehensive use of the knowledge in the various relevant fields and adopts computer technology and simulation technology to detect automobile wheel positioning performance. It is highly practical and can be applied to a wide range of engineering.