A method and system for detecting hidden cracks at the splicing position of a photovoltaic module based on machine vision
By constructing a feature enhancement network at the splicing point and a frequency-spatial dual-branch feature extraction method, the problem of low differentiation between weak features and noise in the detection of microcracks at the splicing point of photovoltaic modules is solved, and efficient and accurate microcrack identification is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- WUXI INSTITUTE OF TECHNOLOGY
- Filing Date
- 2026-05-09
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies struggle to effectively distinguish between subtle microcrack features and background noise in the detection of microcracks at photovoltaic module splicing points, resulting in poor detection accuracy and a high false positive rate.
A machine vision-based method for detecting microcracks at the splicing points of photovoltaic modules is adopted. By constructing a feature enhancement network at the splicing points, the splicing point features are extracted using an edge attention mechanism and adaptive mask constraints. Combined with a weak feature differentiation gain mechanism and frequency-spatial dual-branch feature extraction, the method achieves differentiated processing of microcracks and noise, and identifies microcracks through adaptive segmentation threshold.
It improves the accuracy and reliability of detecting microcracks at the splicing points of photovoltaic modules, reduces the false positive and false negative rates, and enhances the efficiency and accuracy of feature extraction.
Smart Images
Figure CN122415569A_ABST