Non-contact Wheel Alignment Technology Based On Binocular Stereo Vision

With the rapid development of the automotive industry,China's auto output and the amount of ownership continue to increase,higher requirements for automobile detection technology are put forward.Wheel alignment as an important test indicator for vehicle driving performance,taking an important position in vehicle performance testing.In order to meet the needs of the user,the wheel alignment technology of auto is developing towards intelligence and automation.Visual measurement is widely used in robot navigation,product quality detection and other industrial fields because of its advantages of fast speed and non-contact.In this paper,the measurement technology with binocular stereo vision is introduced into the wheel alignment detection.The design of the calibration system for the wheel alignment is completed.We also provide a wheel measurement method for wheel aligner with binocular stereo non-contact,detect the spatial position of the wheel and realize the non-contact measurement of wheel.The main research of this paper is divided into the following aspects:Firstly,we complete the calibration system based on binocular stereo vision for wheel alignment,the systems consists of the monocular camera calibration system and the binocular camera calibration system.Through establishing the model of monocular camera imaging,we analyze the principle of monocular camera imaging and set up a mathematical model for camera calibration.We use the planar template method to obtain the internal matrix A of internal parameters and the rotation matrix r and translation vector t of the external parameters for monocular camera.And then we establish a binocular stereo imaging model.According to the principle of binocular stereo imaging,we calibrate the relative position between the left and right cameras,namely the rotation matrix R and the translation vector T.Secondly,because of the characteristics of the wheel images are not obvious and the texture is repeated,the wheel images can not be matched in directly.Therefore,the method of digital image processing is used to process the image of the wheel images.Through image segmentation,closing operation,removing wheel image boundary and removing small area of wheel images,we can find wheel hub and fit hub edge.Therefore,the measurement with binocular stereo vision can be used to solve wheel alignment problems.Finally,according to the principle of the stereo matching algorithm,we improving the existing stereo matching algorithm and design a new stereo matching algorithm for wheel images.After the wheel images are corrected and fitted,stereo matching is performed according to the features of the wheel hub image obtained after processing.The improved stereo matching algorithm only needs to search the matching points on the same scanning line and transform the two-dimensional search into one-dimension.There is no need to traverse the whole image,so the computing time is greatly saved.According to the characteristics of the wheel,the only matching point is located,and the accuracy of the stereo matching is improved.Finally,according to the matching point,the 3D coordinates of the edge point in the hub are obtained,thus the spatial position of the wheel axle is obtained.