TOFD图像中焊缝埋藏缺陷的智能评定与分级
Intelligent Evaluation and Classification of Embedded Weld Flaws in TOFD Images
余焕伟 1任绪凯 1廖晓平 2欧阳星峰 1杜锡勇1
作者信息
- 1. 绍兴市特种设备检测院 绍兴 312071;绍兴市特种设备智能检测与评价重点实验室 绍兴 312071
- 2. 浙江德力装备有限公司 绍兴 312599
- 折叠
摘要
TOFD检测中缺陷的评定与分级主要依靠人工进行,需要消耗大量的精力和时间,而且还存在误判的可能性.针对该问题,本文提出一套不依赖人工干预的缺陷自动评定与质量分级方法,包括TOFD图像分区与标定、灵敏度评价、缺陷区域分割、评定与分级.试验结果表明,本文方法与传统人工方法相比,全程无须人工干预,能显著提高分析评定人员的工作效率,特别适合大型特种设备制造及安装过程中批量化TOFD图像的缺陷评定与质量分级.
Abstract
In TOFD detection,the quantitative evaluation and classification of flaws mainly carried out manually,consumes a lot of energy and time,and there is the possibility of misjudgment.Aiming at the above problem,a set of automatic defect assessment and quality classification methods which didn't depend on human intervention were proposed,including TOFD image partition,automatic calibration,image sensitivity evaluation,defect region segmentation,evaluation and quality classification.The experimental results showed that the proposed methods didn't need human intervention in the whole process compared to the traditional methods and could significantly improve the processing efficiency of batch TOFD images during the manufacture and installation of large special equipment.
关键词
衍射时差超声检测/焊缝埋藏缺陷/评定与分级/阈值分割Key words
Time of flight diffraction(TOFD)/Embedded weld defects/Evaluation and classification/Threshold segmentation引用本文复制引用
基金项目
浙江省市场监督管理局科技项目(20200333)
浙江省市场监督管理局雏鹰计划培育项目(CY2023215)
中国博士后科学基金(2023M742598)
出版年
2024
NETL
NSTL
万方数据