基于Canny算子的焊缝图像边缘提取技术  

Edge detection of weld image based on Canny operator

作　　者：谢志孟[1] 高向东[1] 

机构地区：[1]广东工业大学机电工程学院,广州510090

出　　处：《焊接学报》2006年第1期29-32,36,共5页Transactions of The China Welding Institution

基　　金：国家自然科学基金资助项目(60375012);广东省自然科学基金资助项目(020176)

摘　　要：焊缝边缘是焊缝图像的最主要特征,选用适当的边缘检测算子得到准确的焊缝信息是焊缝图像处理的关键步骤。选用并研究Canny算子来提取焊缝的边缘特征,详尽地叙述Canny算子的基本原理和算法实现过程。最后为了验证Canny边缘检测算子的效果,分别用Canny、K irsch、Prew itt、Robert、Sobel和Gauss-Lap lace算子对焊缝图像进行边缘检测试验,根据边缘处理结果和焊缝中心坐标图像,得出Canny边缘检测算子是边缘检测的一种十分有效的方法,适合于视觉传感的焊缝跟踪图像处理过程。Weld edge is the most important characteristic in a weld image.The fact of selecting an appropriate edge-detecting operator and obtaining the accurate weld edge information is a key process in weld image processing.The canny operator was used to extract the weld edge,and its principle and implementation were discussed.And the canny operator,kirsch operator,prewitt operator,robert operator,sobel operator and gauss-laplace operator were applied in detecting the weld edges to test the effects,respedively.The analyzsis of the weld image and weld center coordinates showed that the canny operator is a very effective way to detect the weld edge,and also suitable for the seamtracking process based on vision sensors.

关 键 词：CANNY算子 边缘检测 焊缝跟踪 图像处理 

分 类 号：TG402[金属学及工艺—焊接]