cat << 'EOF' > README.md
This project represents the computer vision module (Phase 1) of an autonomous agricultural robot designed for plant disease monitoring. Utilizing a YOLO-based object detection architecture, this model detects plant leaves in complex greenhouse environments with high precision.
It is specifically engineered to overcome common agricultural challenges such as occlusion (overlapping leaves), variable lighting (backlight/shadows), and visual similarities (distinguishing green tomatoes from green leaves).
Current Status: ✅ Phase 1 Completed (Detection) | 🔄 Phase 2 (Disease Classification) In Progress
The model was trained for 120 epochs on approximately 10,000 labeled instances at 1024px input resolution (imgsz=1024).
| Metric | Score | Significance |
|---|---|---|
| mAP@50 | 0.83 | High detection accuracy |
| Precision | 0.83 | Low false positive rate |
| F1-Score | 0.76 | Optimal precision–recall balance |
-
Clone the repository
git clone https://github.com/UmutUsenmez/Autonomous-Plant-Disease-Robot.git cd Autonomous-Plant-Disease-Robot -
Install dependencies
pip install -r requirements.txt -
Run inference
python src/inference.py path/to/your/image.jpg
- Phase 1: Robust Leaf Detection — Completed
- Phase 2: Disease Classification — Completed (https://github.com/UmutUsenmez/Disease-Diagnosis-Module)
- Phase 3: AI-Driven Treatment Recommendation and Reporting
Umut
Mechatronics Engineering Student at Yıldız Technical University
Focus: AI · Computer Vision · Robotics
EOF