主动视觉的焊缝跟踪图像分析

Image analysis of weld tracking based on active vision

焊接结构件应用领域广、需求量大,在提高效率的前提下保证焊接质量是非常必要的。随着人工智能技术的发展,焊接智能化生产成为了可能。文中基于焊接机器人构建视觉焊缝自动跟踪系统,结合工业相机于机器人系统标定技术,通过采用中值滤波平滑去噪、灰度拉伸增强对图像进行预处理。结合Canny算子提取图像边缘,Zhang改进细化算法获取图像单像素条纹,最小二乘法对骨骼线上的点进行拟合,最终得到特征点进行实际坐标转换获取焊缝中心点位置。试验表明,该研究具有一定的可行性,满足实际需求。

Welded structural parts are widely used and in great demand.It is necessary to ensure the welding quality on the premise of improving the efficiency.With the development of artificial intelligence technology,intelligent welding production has become possible.In the paper,a visual weld tracking system based on welding robots was constructed.Combining with the calibration technology of industrial camera in robot system,the image was preprocessed by using median filter to smooth noise and gray stretch enhancement.Combining Canny operator to extract image edge,Zhang improved thinning algorithm to obtain image single-pixel fringes,least squares method to fit points on skeleton line,and finally get feature points for actual coordinate transformation to obtain the position of weld center points.Experiments showed that the research was feasible and met the actual needs.