基于分形和数学形态学的TIG焊熔池图像分析  

Analysis of Molten Pool Image of TIG Welding Based on Fractal and Mathematical Morphology

作　　者：薛家祥[1] 刘晓[1] 朱思君[1] 王振民[1] 

机构地区：[1]华南理工大学机械工程学院,广东广州510640

出　　处：《华南理工大学学报（自然科学版）》2007年第8期7-10,共4页Journal of South China University of Technology(Natural Science Edition)

基　　金：广东省自然科学基金资助项目(04020100)

摘　　要：焊接图像中不可避免地存在烟尘和弧光等干扰造成的噪声,而传统的图像处理方法对噪声非常敏感,因此很难准确地对焊接图像进行处理.文中提出了一种基于分形和数学形态学理论的综合图像处理方法.该方法利用分形理论中的离散分数布朗随机场理论得到按分形维分布的熔池灰度图像,然后利用二值形态学中的连通域检测去除噪声并得到目标图像,进而利用形态学中的开运算和闭运算得到图像的强连通边界.试验结果表明,该方法能够快速、准确地检测熔池的边缘、高度、宽度以及面积等信息.The processing of welding images is unavoidably accompanied with noises due to the dust, strong arc light and other interference factors in the welding process. As the traditional methods of image processing are sensitive to the noise, it is difficult to accurately process the image. In order to solve this problem, this paper presents a novel composite image processing method based on the fractal and mathematical morphology theories. In this method, first, the gray-scale image of the molten pool is obtained according to the fractal-based discrete Brown random field theory. Next, the connected domain is detected based on the binary morphology theory and is then used to remove the noise and obtain the target image. Moreover, the successive edges of the image are obtained by means of the Open-Close operation on the basis of the mathematical morphology theory. Experimental results show that the proposed method is of high speed and accuracy in the recognition of molten pool information including the edge, the height, the width and the area.

关 键 词：TIG焊 熔池 图像 分形维数 数学形态学 边缘检测 

分 类 号：TG403[金属学及工艺—焊接]