管道焊缝探伤机器人的视觉牵引算法

Vision-based navigation algorithms for pipeline weld inspection robot

设计了一套管道焊缝探伤机器人的视觉牵引算法,用于完成高性能的焊缝自动定位任务。机器视觉系统采集管道图像后,以焊缝形态为特征,先对图像进行形态学滤波,然后用自适应阈值进行图像分割,最后用Hough变换从图像中准确检测出焊缝相对位置。试验结果表明,整套算法可靠、准确、实时性好,完全满足管道探伤机器人焊缝自动定位系统的性能要求。

In order to locate the weld effectively and automatically, a set of vision-based navigation algorithms for pipeline weld inspection robot is proposed in the paper. After the pipeline image is acquired by visual system, it is firstly pretreated with a morphology filter, and then segmented by an self-adaptive threshold, finally the weld is located accurately by Hough transform. The result shows that the algorithms are reliable, accurate and real-time, thus completely meeting the requirement of the weld locating system.