结合Sobel算子与Snake模型的焊缝图像边缘提取  

Seam Extraction Method Based on Sobel Operator and Greedy Snake Model

作　　者：李静[1,2] 李芳 秦小麟[1] 朱伟 张敏[1] 钱鲁泓 

机构地区：[1]南京航空航天大学信息科学与技术学院,南京210016 [2]昆山工研院工业机器人研究所,江苏昆山215301

出　　处：《小型微型计算机系统》2011年第12期2491-2494,共4页Journal of Chinese Computer Systems

基　　金：江苏省科技成果转化专项基金项目(BA2007058)资助

摘　　要：随着焊接自动化、智能化研究的深入,人们越来越注重焊接质量的自动控制.焊接过程中焊缝始终保持对中间隙中心是实现焊接过程自动化的前提条件.在管道打底MIG焊视觉跟踪系统中,首先基于光学成像的几何原理,通过CCD摄像机获取较清晰焊缝图像,使用Sobel变换获得Snake模型的初始轮廓点,然后使用贪婪Snake模型得到焊缝的准确边缘.实验结果表明,与原有Sobel变换算法相比,该方法显著提高焊缝提取的精度.With the development of welding automation and intelligence,It′s very important to accurately extract the edge of welding seam for automatic welding quality control.According to the characteristics of MIG welding,an image sensing system of the welding seam was established.Based on a clearer weld pool image,A new method combined Sobel operator and greedy Snake active model to extract the edge of welding seam was presented.The experimental results show that this method can obtain a more accurate seam edge comparing with Sobel operator method.

关 键 词：焊缝跟踪 机器视觉 Snake主动轮廓模型 MIG焊 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]