摘要
对射线实时成像检测技术的发展现状进行了分析 ,指出了目前存在的主要问题。以焊接缺陷的射线实时成像检测为对象 ,简要介绍了图像增强及去噪声等预处理方法。对缺陷提取技术进行了研究 ,设计了基于灰度梯度的焊缝边界提取方法 ,采用基于分离焊缝的拟合 DSA方法进行特征提取 ,详细介绍了有关算法并给出了实际处理结果。另外 ,提出了提取结果相关性修正的概念及算法并对缺陷类型识别方法进行了研究。
The technical level of present real time radiography inspection system was analyzed with the evaluation of real time radiography techniques for weld defects. Image preprocessing methods were introduced for image enhancement and noise elimination. Defect extraction techniques were studied with an effective weld boundary extracting algorithm developed based on gray level. A DSA method based on background fitting for isolated weld seams was proposed for feature extraction with the concept of correlativity modification put forward for the extraction results. Automatic recognition techniques for weld defects were also studied. These techniques were then combined into a real time radiographic inspection system for weld defects.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2001年第2期25-29,共5页
Journal of Tsinghua University(Science and Technology)
