It's necessary for automobile to detect and adjust four-wheel alignment parameters regularly, due to the significant effect on improving stability, enhancing security and reducing tire wear of automobiles. In order to measure the parameters that determined by relative position and posture of four wheels to the automobile cab, this paper proposes a method which applies monocular vision of linear structure light to wheel pose measurement. Firstly, space coordinates of feature point cloud are calculated out from the principle of structured light. Then, an algorithm is designed to determine the normal vector of wheel tangent plane and measure the wheel pose. Finally, actual experiments that by evaluation of adjusted wheel angle measurement are carried out to verify the system accuracy. The corresponding studies can be applied in designing and developing 3D four-wheel alignment system that based on structured light.
Aiming at the problems of existing crack extraction algorithms which are difficult to achieve fast and accurate crack extraction of image, an algorithm of crack detection based on Median Filter and Hessian Matrix is proposed. Firstly, median filter of crack gray image in 4 directions, Level, 45 degree, vertical and -45 degree, is conducted, by which noises are removed and roughly extracted crack is obtained. Then according to the Hessian matrix feature of extracting image linear feature, convolution of Differential operation of the Hessian matrix is adopted, and crack is further extracted through eigenvalues response and changing standard deviation of Gaussian function. The proposed algorithm validity is verified by comparison with other crack extraction algorithm. The results show that this algorithm has obvious accuracy rate in crack extraction.
In order to solve the robustness and efficiency problems of chessboard corner detection under on-site condition, a method based on the square-closed loop template and local gray symmetry factor is proposed for detecting sub-pixel chessboard corners with high precision and efficiency. Interest points on edges of original image is detected by square-closed loop template according to the transition times on the template, on this basis, corners are roughly detected by averaging the adjacent coordinates of interest points; gray symmetry factors, calculated in the local neighborhood of roughly detected corners, are used as the weighting factors to precisely detect corners on sub-pixel level. Experimental results indicate that computing speed and positioning accuracy of this method was obviously increased. The corner detection performance could be significantly improved using the proposed method.