A novel laser line detection algorithm for robot application

摘要

In this paper, we proposed a novel laser line detection method for robot applications using laser scan line. The method is a robust algorithm dealing with many challenges of line detection, particularly in laser line detection, such as the saturation phenomenon of laser light, the effects of white ambient light and the segmentation of laser scan line caused by obstacles. First, the method takes advantages of YCbCr color space in order to enhance laser signal and reduce the effects of white ambient light for distinguishing as well as increasing the accuracy of line detection algorithm. After that, it automatically estimates the saturation of laser light and adjusts the exposure by capturing a sequence of images with different exposures. Finally, a robust line segment extraction algorithm is applied to extract the real laser line segments and form the detected laser scan line. The experiment results with various scenes and lighting conditions show the accuracy and robustness of the proposed method.