基于有监督过渡区的焊缝X射线图像分割  

Segmentation of weld seam in X-ray images based on supervised transition region extraction

作　　者：佟彤[1] 蔡艳[1] 孙大为[1] 吴毅雄[1] 

机构地区：[1]上海交通大学激光制造与材料改性重点实验室,上海200240

出　　处：《焊接学报》2014年第3期101-104,118,共4页Transactions of The China Welding Institution

摘　　要：针对X射线焊缝检测图像中存在大量背景冗余信息,焊缝和缺陷难于准确检测提取的问题,提出一种基于先验知识的有监督过渡区域提取及阈值分割方法.根据焊接图像本身的特点,通过先验知识对样本图像进行训练,确定某个区间来估算图像过渡区域的灰度范围,按照模糊子集理论,给出一种新的加权算子来描述局部窗口内灰度级的变化,从而能充分考虑到局部窗口内灰度级变化的频率和幅度,通过计算过渡区域像素的灰度均值,将其作为阈值对图像进行分割.结果表明,该方法能准确地将目标缺陷从焊缝X射线图像中分割出来,具有良好的适应性.Regarding that the traditional algorithm can only achieve a low successful defect segmentation result for the redundant background information in X-ray image,an efficient thresholding approach to confine transition region of the object based on supervision is introduced. The methods first implement gray level range-estimation,using sample images in the light of priori knowledge. According to the theory of fuzzy subset,a new descriptor taking both frequency and degree of gray level changes into account is developed. The final segmentation threshold can be determined by mean of the histogram of the transition region.The proposed method was compared with their counterparts on a variety of images,and the experimental results show its effectiveness.

关 键 词：无损检测 焊缝提取 先验知识 过渡区 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]