基于高频加强变换的焊接缺陷检测方法

Welding defect detection method based on high frequency enhancement of transform

根据X射线线阵实时成像检测压力容器的特点,提出焊缝图像处理及快速发现焊接缺陷的方法。利用自适应中值滤波方法和类间、类内方差比分割法及数学形态学方法对原始焊缝图像进行降噪和提取焊缝区域,并对提取的焊缝内部进行频域锐化的高频加强变换.仿真结果表明,自适应中值滤波能够有效去除噪声,并保持焊缝和缺陷的边缘细节;类间、类内方差比分割方法增强了图像目标与背景的对比,与数学形态学方法并用能准确地提取焊缝区域;高频加强变换能够为焊缝缺陷提供清晰的视觉效果,从而根据其显示的特点对焊接缺陷进行提取.

According to the characteristics of examination of pressure vessel with X-ray linear array realtime i/naging, a method for welded seam image processing and the fast discovery of welding defect was proposed. The methods of adaptive median filter, variance ratio of interclass and intraclass segmentation, and mathematical morphology were used to decrease the noise of original welded seam image, to extract welded seam region, and to carry on high frequency enhancement transformation for the interior of extracted welded seam with frequency range sharpening. The simulation result indicated that the adaptive median filter could effectively remove the noise, and maintain the details of welded seam and defection edge, and the variance ratio of interclass and intraclass segmentation could strengthen the image objective and background contrast. When the latter was used together with the method of mathematic morphology, the weld region would accurately be extracted. The high frequency enhancement transformation could provide a clear visual effect for weld defection, so that the welding defect could be extracted according to its demonstrated characteristics.