柱上设备带电裸露点绝缘化判定方法及系统
By using illumination and image capture modes based on light intensity determination, combined with a multi-feature linear weighted fusion strategy and a deep learning model, the problems of poor image acquisition quality and inaccurate defect identification in the insulation retrofit of live exposed points of pole-mounted equipment are solved, and high-precision defect detection is achieved under different lighting conditions and complex layouts.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- STATE GRID INTELLIGENCE TECHNOLOGY CO LTD
- Filing Date
- 2025-08-15
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies make it difficult to achieve high-quality image acquisition under different lighting conditions and complex equipment layouts during the acceptance of insulation retrofits for exposed live points of pole-mounted equipment. This leads to inaccurate defect identification, and the shaking of the insulating rod during manual operation affects image quality.
We employ a lighting and image capture mode based on illumination intensity determination, combined with a focal length adjustment method using a multi-feature linear weighted fusion strategy. We utilize YOLOv5 and YOLOv8 deep learning models for image quality assessment and anomaly/defect detection. By weighted summation of classification loss, bounding box regression loss, and confidence loss, we achieve automatic image quality adjustment and defect identification.
It improves image quality under different lighting conditions and complex equipment layouts, accurately identifies and classifies different defect features, improves detection accuracy, and solves the problems of poor image quality and inaccurate defect identification.
Smart Images
Figure CN120992625B_ABST