Commercial product
Solar Panel Quality Control
Oct 2024 – Dec 2024
Eliminated missed-cell errors in cell counting and improved small-object detection with SAHI. Delivered as a subscription-based, server-licensed commercial product.
- 95%+small-object detection accuracy
A quality control system that detects and evaluates each solar panel cell individually on the production line.
- Redesigned the cell-counting logic by combining image processing with geometric estimation, eliminating the missed-cell errors of the previous version.
- Integrated SAHI to improve small-object detection, raising model accuracy to 95%+.
- Rebuilt the desktop interface from scratch with database integration.
- Developed a server-validated licensing system enabling subscription-based commercial delivery.
Stack
- Python
- OpenCV
- YOLOv8
- SAHI
- PyQt5
- MongoDB
- SQLite
- FastAPI