计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (33): 237-239.

• 工程与应用 • 上一篇    下一篇

薄壁焊缝X射线图像缺陷的自动提取与分割

王明泉1,2,杨 静2,李志刚1,王 玉1   

  1. 1.中北大学 仪器科学与动态测试教育部重点实验室,太原 030051
    2.清华大学 工程物理系,北京 100084
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-11-21 发布日期:2007-11-21
  • 通讯作者: 王明泉

Automatic defect extraction and segmentation in X-ray images of welding seam

WANG Ming-quan1,2,YANG Jing2,LI Zhi-gang1,WANG Yu1   

  1. 1.The Ministry Education Key Lab for Instrumentation Science and Dynamic Test,North University of China,Taiyuan 030051,China
    2.Department of Engineering Physics,Tsinghua University,Beijing 100084,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-11-21 Published:2007-11-21
  • Contact: WANG Ming-quan

摘要: 根据薄壁焊缝X射线图像的特点,针对基于数学形态法的图像分割技术进行了改进。首先利用数学形态学选取适当的结构元素模拟图像背景,然后利用数字剪影法提取缺陷目标。文章将使用较好的几种阈值化方法进行分析,比较各种方法的基本思想、优缺点及使用范围,最终提出选择迭代分割方法得到图像最佳阈值将图像二值化。实验结果表明,针对不同的缺陷均能得到轮廓清晰的分割效果,为缺陷特征参数的提取和识别的实现打下坚实的基础。

关键词: 图像分割, 数学形态学, 迭代阈值

Abstract: Regarding the characteristic of X—ray detection images of thin welding,image segmentation was improved based on mathematical morphology in this paper.First of all,background can be simulated by proper element in mathematical morphology,and then defect region were extracted successfully using arithmetic of digital subtraction.The present paper analyses a few methods for threshold selection.And the basis ideas of various methods as well as their advantages and defects are critically reviewed.The experimental results indicate that both of methods can accomplish defect extraction and segmentation automatically,which will lay a good foundation for flaw feature parameter extraction and recognition.

Key words: image segmentation, mathematical morphology, iterative threshold