摘要
焊接是钢结构生产的主要方法。射线检测方法是检查焊接缺陷的几种主要方法之一。射线检测结果的评定方法有两种:人工评定和计算机辅助评定。人工评定检测结果时,评定人员工作量大,评定结果易受评定人员主观因素影响。采用计算机对X射线检测结果进行分析和识别,可以大大提高工作效率。分析识别的基础是图像的分割。本文分析Mumford-Shah模型和主动轮廓的CV模型,并将其首次应用到焊缝缺陷的分割之中。目前多数文献选择单个初始圆为曲线演化的初始轮廓,本文将初始轮廓选择为多个初始圆,此方法极大加快了曲线演化的速度。
Weld is the main way of constructing steel structure. X ray detection is one of the few major techniques in detecting weld defects. There are two methods for evaluating the results of X ray detection： manually and automatically evaluation. In a manual way, the person may be tired so the result is not reliable, and the evaluation is subjective. The computer aided analysis and recognition can improve the efficiency. The basis of analysis and recognition is image segmentation. This article is based on CV model which is an development of Mumford-Shah model and active contour by Tony, Chan and Luminita. And we apply this model in welding seam defects detection.
出处
《CT理论与应用研究（中英文）》
2008年第4期55-61,共7页
Computerized Tomography Theory and Applications
关键词
焊接
缺陷
分割
主动轮廓
weld defect segmentation active contour
