线结构光视觉传感焊缝跟踪图像处理

Image Processing of Line Structured Light Vision Sensing Seam Tracking

提出了一种基于灰度形态学的图像处理算法,有效去除了线结构光视觉传感焊缝跟踪图像中飞溅、弧光和烟尘等干扰。根据飞溅在图像中表现为明暗相间的条纹的特点,采用灰度形态学开操作,通过合理设计结构元素,去除了飞溅干扰。针对不同图像受弧光、烟尘影响不同,背景灰度值差异较大,不利于二值化阈值选取的问题,采用图像相减和顶帽变换消除其影响。最后,采用动态开小窗与最小二乘法相结合的方法准确提取到了焊缝特征点。该算法能有效去除飞溅、烟尘等干扰,实时性好。

A new seam image processing algorithrn based on grayscale morphology was proposed, which could effectively remove the splash, dust and other interference in the line structured light vision sensing seam tracking image. According to the characteristics of the splash, through reasonablly designing structure elements, the grayscale morphological opening was used to remove the interference of the splash. Because of different images caused by arc light and dust, the influences were different, which was not conducive to binary threshold selection. By using image subtraction and top-hat transformation, its influence was eliminated. Finally, by using the method of dynamic ROI and the least square method, the feature points of the weld seam were extracted accurately. This algorithm can effectively remove the interference of the splash and dust, and has a good real-time behavior.