灰度形态学提取焊接熔池图像边缘技术

Extracting borderline of CO₂ arc welding molten pool image using gray-scale morphology

摘要: 在由CCD捕获的CO₂焊接熔池图像中,存在着大量的点状噪声和线状噪声,要精确地检测出熔池图像的边缘比较困难。采用基于数学形态学理论中的灰度形态学数字图像处理技术,结合焊接熔池图像自身的特点,借助灰度图像二值化取阈值法中的矩量保持法确定结构元素值的大小,通过合理设计形态学滤波器形状和运算顺序,成功地提取出了焊接熔池图像的边缘,通过比较,新的图像处理算法的处理效果非常明显,对后续的焊接质量控制和焊缝跟踪工作提供了非常有用的熔池信息。

Abstract: For CO₂ arc welding molten pool image, there are too many point and line noises, so it is difficulty to accurately extract borderline of molten pool image.We made use of the image processing technology of gray-scale morphology based on mathematical morphology and design the filter figure and operation order by means of moments-preserving method to extract successfully borderline of molten pool image of CO₂ arc welding.So the processed image provides useful molten pool information for welding process control and welding quality control.