激光视觉焊缝跟踪图像处理与坡口识别

摘要

针对激光引导焊缝跟踪过程中坡口图像特点,首先对图像进行Radon变换,检测出激光条纹所在位置;然后对变换后图像进行处理,消除噪声对应的信息,保留激光条纹所对应的信息,再进行Radon逆变换复原图像,可较好地消除噪声;最后,采用OTSU法对图像进行分割,并识别出坡口区域.该方法效果较好,抗干扰性较强,耗时较少,适合于激光视觉焊缝跟踪过程.%According to the image characteristics in weld seam tracking based on laser vision, Radon transform is used at first to find the location of laser stripe; then, the transformed image is processed by using the method that the information coming from laser stripe is kept and other information coming from noises is removed. After that the inverse Radon transform is carried out to recover the original image. Finally, OTSU method is used for image segmentation so that the groove region can be recognized. These methods have the merits of good effects, less time cost, and the strong antidisturbance performance, therefore they are suitable for seam tracking based on laser vision.