焊缝起始点的识别是实现机器人智能焊接系统的一个重要环节。利用多结构元素的形态学方法对CCD在自然光条件下采集到的直焊缝图像进行处理,提取直焊缝边缘后借助于Hough变换得到焊缝初始点位置。试验结果表明,与传统的边缘检测方法和单结构元素形态学方法相比,本文提出的多结构元素形态学边缘检测方法边缘检测效果更佳,能有效去除噪声的同时,保留更多的边缘细节信息,为后续的图像处理操作奠定了基础,是一种有效的边缘检测方法。
The recognition and positioning of initial welding position is an important part of intelligent robot welding system. The multi-structure elements morphology was used to process the straight seam image which was collected by CCD under nature light condition. The position of initial welding gained by transformation after straight seam edge was abstracted. The experiment results show that compared with the tradition edge detection method and single structure element morphological method, the proposed algorithm has better edge detection. The method can not only decrease the noise effectively, but also reserve more detail of edges, which can lay the foundation for the subsequent image processing operation.