Sugarbeet Weed Segmentation
Semantic segmentation for agricultural weed detection in sugarbeet fields
Quick Summary
- Developing real-time weed detection system for precision agriculture on Farm-ng Amiga robotic platform
- Deployed PyTorch semantic segmentation model (ERFNet) on NVIDIA Jetson Xavier via TensorRT FP16
- Achieved 8.6x inference speedup (718ms → 84ms) enabling real-time field operation
- Trained and evaluated on PhenoBench dataset for sugarbeet/weed discrimination
Tools: PyTorch, TensorRT, NVIDIA Jetson, OpenCV, Python
For more information, visit: https://github.com/gabe-zhang/sugarbeet-weed-segmentation