针对射线检测焊接图像内缺陷的提取,本文给出了一种利用小波分析对焊接图像中列灰度曲线奇异点检测进行缺陷提取的方法。把每一列灰度值曲线的微分曲线作为被分析函数f(x),根据Lipschitz指数与小波变换模极大值的关系,利用小波模极大值处检测出曲线f(x)上的所有奇异点,并利用Lipschitz指数来估算边缘点奇异程度,由此实现焊缝和缺陷边缘的精确检测。仿真结果表明,利用小波分析方法可取得较好的焊缝、缺陷边缘提取效果。 In this paper, a method of extracting defects by using wavelet analysis to detect the singularity of column grayscale curve in welding image is given in this paper. Taking the differential curve of each gray value curve as the analyzed function f (x), all the singularities on the curve f (x) are detected by using the relation between the Lipschitz index and the maximum of the wavelet transform module. Point, and using the Lipschitz index to estimate the singularity of the edge point, thus achieving the accurate detection of the weld and the edge of the defect. The simulation results show that the wavelet analysis method can get a better edge of the weld and defect extraction.