.png){kind=icon}
返回
Vision-based quality evaluation method towards automated penetrant testing
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ndteint.2025.103334.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Penetrant Testing (PT) is one of the most widely utilized Non-Destructive Testing (NDT) methods; however, it currently relies on manual operation, which presents various challenges. Although initiatives to develop robotic PT systems have been undertaken, full automation remains unachieved due to the absence of established evaluation methods. In this paper, we propose a vision-based quality evaluation method aimed at automating PT, addressing the existing qualification criteria that rely on visual observation. In particular, we firstly introduce a detection network designed to process images and evaluate the quality of the penetrant process effectively. Moreover, since the PT process potentially suffers from poor lighting, we develop an image preprocessing network to improve the performance of the penetrant evaluation. In addition, we construct a dataset containing annotated images of the penetrant procedure, which is used to train the proposed network and conduct experiments. Our results, both quantitative and qualitative, demonstrate that the proposed network exhibits remarkable precision and robustness in penetrant evaluation tasks.
Keyword:
Penetrant testing
Deep learning
Detection
Quality evaluation
Low-light enhancement
期刊
N
IF:
4.5
论文数:
3.1K
被引数:
9.0K
机构
引用论文
Increased Pathogenicity and Shedding in Chickens of a Wild Bird–Origin Low Pathogenicity Avian Influenza Virus of the H7N3 Subtype Following MultipleIn VivoPassages in Quail and TurkeyH7N3亚型野鸟源低致病性禽流感病毒在鹌鹑和火鸡中经过多次体内传代后，在鸡中表现出致病性和排毒增强。
PREAI
分享收藏分享收藏Can crayfish take the heat?Procambarus clarkiishow nociceptive behaviour to high temperature stimuli, but not low temperature or chemical stimuli.png)
Biology Open
OAAI
Biology OpenIF0
OAAI分享收藏
查看更多内容