基于形态学算法的焊缝图像处理技术  

Image processing technology of weld seam based on morphology algorithm

作　　者：侯琳 张海传[1] HOU Lin;ZHANG Haichuan(School of Information Science and Engineering, Dalian Polytechnic University, Dalian 116034, China)

机构地区：[1]大连工业大学信息科学与工程学院,辽宁大连116034

出　　处：《大连工业大学学报》2022年第2期142-146,共5页Journal of Dalian Polytechnic University

基　　金：辽宁省教育厅科学研究项目(J2019022).

摘　　要：焊接作为工业自动化领域中重要的一环,其结果的好坏直接决定了整个生产过程效率的高低。由于焊接过程会受到弧光、烟尘等因素的干扰,为了得到清晰准确的焊缝图像,首先采用了灰度化和中值滤波对图像进行预处理,然后采用最大类间方差法对图像做二值化,达到减小噪声干扰的目的。再采用改进的形态学方法,运用多种不同的结构元素对图像做处理,保留图像中有用的分量,最后利用最小二乘法进行焊缝直线拟合,得到精确的焊缝位置图像。实验结果表明,该方法原理简单,快速有效地得到了较为清晰的焊缝图像,便于之后的焊缝定位。Welding is an important part in the field of industrial automation,and the quality of the result directly determines the efficiency of the whole production process.The welding process will be disturbed by arc light,smoke and other factors.In order to get a clear and accurate image of welding seam,the grayscale and median filter were used to preprocess the image,and then the maximum inter-class variance method was used to binarize the image,so as to reduce the noise interference.Then,the improved morphological method was used to process the image by using a variety of different structural elements to retain the useful components in the image.The least-square method was used to carry out linear fitting of the weld,and the accurate image of the weld position was obtained.The test results show that the principle of this method is simple,and the relatively clear weld image is obtained quickly and effectively,which is convenient for subsequent weld positioning.

关 键 词：图像处理 焊缝检测 形态学处理 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]