面向深地医学救援的深度引导图像去雾方法与系统
By introducing scale-coordinated attention and error-aware guidance modules into the u-net network, the problem of image quality degradation in deep-earth medical rescue was solved, multi-scale feature co-optimization was achieved, and the accuracy and robustness of image monitoring and detection systems were improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- YUNLONG LAKE LAB OF DEEP UNDERGROUND SCI & ENG
- Filing Date
- 2025-11-11
- Publication Date
- 2026-07-17
AI Technical Summary
In deep-earth medical rescue scenarios, there are serious problems of image quality degradation, including low contrast, local overexposure, local fog residue, and reduced visibility. Existing methods suffer from scale confusion and limited scale coordination capabilities in multi-scale feature extraction and feature fusion, which affects the accuracy of image monitoring and detection systems.
A depth-guided image dehazing method is adopted, which incorporates a scale-coordinated attention module and an error-aware guidance module into the u-net network architecture. By combining cross-scale tensor encoding, multi-scale encoding, and error-aware guidance mechanisms, the interaction and optimization of multi-scale features are realized, thereby enhancing the collaborative performance of image dehazing and depth estimation.
It improves the accuracy and robustness of image dehazing and depth estimation, enhances the accuracy of monitoring camera systems and intelligent detection systems in mines, and strengthens image restoration in non-uniform fog environments.
Smart Images
Figure CN121304493B_ABST