基于幂函数标准化系数的激光光条中心提取方法

In the vision measurement systems of structured light for automotive morphology, the key point of detection accuracy is accurately determining the central position of the laser stripe in the image. An active vision method for extraction of laser line center based on standardized coefficient of power function is proposed. First, preprocessing of image using difference image method is conducted to realize image segmentation of the laser stripe; Secondly, pixel coordinates of the feature points to be selected of the light stripe center are obtained using Hessian matrix, and two eigenvalues and eigenvectors of the Hessian matrix of feature points to be selected of the light stripe center are obtained; Finally, according to the distribution characteristics of the grayscale of the laser light center in the image, a determining normalization method for eigenvalue based on power function is proposed, extraction of the center of the laser stripe in sub-pixel image is realized.