基于深度学习的核安全壳缺陷检测定位方法及系统
By using remote acquisition equipment and deep learning models to automatically detect defects in the nuclear containment structure, the problems of long time consumption and high risk associated with traditional manual inspection have been solved. This enables rapid and accurate detection and visual location of various defects, thereby improving the safety and operational efficiency of nuclear power plants.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- WUHAN UNIV
- Filing Date
- 2023-11-02
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional manual visual inspection of nuclear power plant containment defects is time-consuming, inefficient, and poses safety risks. Existing deep learning-based inspection methods mainly focus on crack detection, failing to fully cover defects such as exposed rebar and repairs.
Remote acquisition equipment is used to automatically acquire images of the nuclear containment structure, construct a containment defect detection model, combine YOLO v5 and Focal NeXt modules, add an attention mechanism, optimize the loss function, and achieve rapid and accurate detection of various defects, and generate a panoramic unfolded image of the containment structure for visual localization.
It enables comprehensive, efficient and accurate inspection of the nuclear containment vessel, reducing inspection costs and risks, and improving the operational efficiency and safety of nuclear power plants.
Smart Images
Figure CN117670795B_ABST