基于形态学多尺度多结构的熔池图像边缘检测  

Edge Detection of Molten Pool Image Based on Morphology Multi-scale and Multi-structuring Elements

作　　者：刘晓刚[1,2] 闫红方 张荣 LIU Xiaogang;YAN Hongfang;ZHANG Rong(School of Mechanical Engineering,Guangxi University of Science and Technology,Liuzhou 545000,China;School of Mechanical Engineering,Guilin University of Aerospace Technology,Guilin 541000,China)

机构地区：[1]广西科技大学机械工程学院,广西柳州545000 [2]桂林航天工业学院机械工程学院,广西桂林541000

出　　处：《热加工工艺》2019年第5期216-219,共4页Hot Working Technology

基　　金：广西自然科学基金资助项目(2014GXNSFAA118310)

摘　　要：熔池图像的边缘提取是实现焊接自动化的关键环节。本文提出采用多尺度多结构的形态学算子对熔池图像进行处理。运用matlab软件进行边缘提取并进行结果验证,同时将其与传统的边缘检测算子及采用不同结构元素的形态学算子进行对比。结果表明,该算子不仅能够提取完整的边缘,而且实现了边缘精准定位和去噪能力的最佳融合。The edge detection of molten pool image is the key to realize the automation of welding.The morphology algorithm based on multi-scale and multi-structuring elements was adopted to process the molten pool image.The edge was extracted by using matlab software and the result was verified.The edge extraction result by matlab was compared with the traditional edge detection algorithms result and morphological algorithms result using different structural elements.The results show that the algorithm can not only exact the complete edge of molten pool image,but also achieve the best combination of precise edge location and denoising ability.

关 键 词：形态学 边缘提取 多尺度 多结构 熔池图像 

分 类 号：TG409[金属学及工艺—焊接]