机器人焊接轨迹自动跟踪视觉技术研究

Research on visual technology of trajectory tracking for robotic welding

针对目前机器人焊接常用的自动跟踪视觉技术方法之不足,即传统的主动视觉方法难于识别空间几何特征不明显的结构,而被动视觉方法又过于依赖灰度梯度特征,在阐述分析焊缝跟踪视觉技术实际应用情况的基础上,提出了利用焊缝纹理特征的视觉检测新思路,给出了基于纹理特征匹配和纹理分割的二种实现方法,并就具体的焊缝图像分析算法的可行性和适用性进行了研究和验证。所提出的在线视觉技术方法为更好地实现机器人多层多道电弧焊接轨迹自动跟踪提供了新的途径。

Aim at seam tracking methods for robotic welding with their limits and disadvantages in applications, that is the conventional active vision methods have difficulties to recognize structure without obvious geometric feature.Neither are the passive vision methods appropriate because of tense dependence on grey-level gradient characteristics.According to the analysis of current application of seam tracking technology, two methods based on the approach of using grain features are proposed ,including the grain feature matching method and the grain segmentation method.Experiments have proved the feasibility and the adaptability of the generated seam image processing algorithm.The presented methods provide new possible solutions for seam tracking of multi-pass/multi-layer robotic are welding.