基于深度学习的核电站混凝土构筑物结构损伤检测研究

打开文本图片集

中图分类号：TQ178文献标志码：A

文章编号：1001-5922（2025）06-0160-04

Research on structural damage detection of concrete structures in nuclear power plants based on deep learning

MA Jun，LI Yi，XU Chao，XU Shichuan，TANG Zhijie（Suzhou Nuclear Power Research Institute Co.，Ltd.，Suzhou ，Jiangsu China）

Abstract：The integrity inspection of concrete structures in nuclear power plant is very important for safe opera⁃ tion. Conventional manual inspection methods have problems such as low efficiency， inaccessibility of some areas， and incomplete data. In this study， based on the classification image recognition technology of deep learning，com⁃ bined with the four-rotor UAV，the automatic identification and data analysis system of surface defects of concrete structures in nuclear power plants was studied and implemented. The experimental results show that the system can accurately identify the common defects on the concrete surface， and can quickly check the concrete structure， which provides a reference for the long-term maintenance strategy of nuclear power plant safety.

Key words：nuclear power plant；concrete；drone；image recognition量缺陷尺寸，然后根据安全标准评估结构状况。（剩余6068字）