CropEye AI
Python
YOLOv8
NVIDIA Jetson
MIT License
AI-powered drone footage analysis for smart farming, with focus on durian orchards
Welcome to CropEye AI, an open-source project designed to bring precision agriculture to life through AI-powered aerial analysis. We run a durian farm in the Southern Philippines, and we leveraged NVIDIA Jetson, CUDA, and YOLO to detect weeds and crop diseases in real-time—all from a drone. Now, we're excited to share our work with the world!
Features
🎯 Real-Time Detection
Catch issues early—like weeds, fungal infections, and dry patches—using a Jetson-equipped drone.
⚡ Edge Computing
No need for constant internet access—perfect for remote farms.
🎯 Precision Targeting
GPS-tagged alerts for precise intervention planning.
🔄 Extensible
Modular architecture supporting custom models and different detection frameworks.
Architecture
🚁 Drone Flight
|
🎥 Live Video Capture
|
🧠 AI Inference (Jetson)
|
Weed/Disease Alerts
|
GPS Tagging & Logging
Getting Started
Prerequisites
- Python 3.8+
- NVIDIA Jetson device (optional, for edge deployment)
- CUDA-compatible GPU (for training)
Installation
git clone https://github.com/patrickrkahn/cropeye-ai.git
cd cropeye-ai
pip install -r requirements.txt
Quick Start
streamlit run app.py
CLI Usage
python smart_farming.py --source farm_footage.mp4
Model Training
yolo train data=durian_farm.yaml model=yolov8n.pt epochs=50
Jetson Optimization
yolo export model=best.pt format=tensorrt
Real-World Deployment
Location: Southern Philippines
Objectives: Early fungal detection, weed control, and overall orchard health monitoring.
Hardware Setup
- Drone with NVIDIA Jetson module
- YOLO-based detection model
- Edge inference for real-time bounding boxes
Results
- Reduced Manual Scouting: Quicker, more frequent scans of the orchard.
- Targeted Chemical Use: Apply fungicides or herbicides exactly where needed.
- Healthier Trees & Higher Yields: Early detection prevents disease spread and fruit loss.
Contributing
We welcome contributions from developers, farmers, and researchers worldwide:
- Fork the repository
- Create a feature branch
- Submit a pull request
Please read our contributing guidelines for more details.
License
CropEye AI is released under the MIT License. See the LICENSE file for details.
"We built CropEye AI in our own durian orchard to detect problems faster and reduce manual labor. Now, we can't wait to see how other farms benefit from it, too!"