Method for extracting ground feature contour and correcting geometric constraint of remote sensing image arranged by photovoltaic
By combining orientation calibration and scale calibration with human-machine collaborative correction, the problem of insufficient contour extraction accuracy in photovoltaic module layout was solved, achieving high-precision photovoltaic module layout and intelligent optimization.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 常州常供电力设计院有限公司
- Filing Date
- 2026-04-28
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for photovoltaic arrays suffer from insufficient contour extraction accuracy, lack of real-time human intervention and model iteration mechanisms, resulting in AI recognition results that cannot be directly used for engineering quantity calculations, and lack of centimeter-level precision in physical dimensions and orientation alignment.
Remote sensing images are acquired through orientation and scale calibration, and then combined with human-machine collaborative correction. Geometric regularization and pixel-level segmentation are performed to achieve a closed-loop process of automatic extraction and manual correction, eliminating pixel-level distortion and achieving centimeter-level accuracy.
It achieves high-precision photovoltaic module arrangement, solves the contour error problem, reduces the amount of manual correction, and improves the level of intelligence through closed-loop optimization.
Smart Images
Figure CN122416262A_ABST