🌿 Plant Disease Detection & Severity Estimation

An intelligent plant disease detection pipeline using YOLOv8 for leaf detection, MobileNetV2 for disease classification, and YOLOv8-seg for accurate severity estimation with spray recommendations.

---

Hugging Face Deployed

🔄 Pipeline Architecture

---

🎯 Severity Estimation

Heuristic vs Segmentation Model

---

✨ Features

Feature	Description
🔍 Leaf Detection	YOLOv8 detects individual leaves
🏷️ Disease Classification	15 disease classes (Tomato, Potato, Pepper)
🔬 Segmentation Severity	YOLOv8-seg for accurate (~85-95%) severity
📊 Heuristic Fallback	Color-based analysis if no seg model
🎯 Spray Tiers	NO_ACTION → LOW → MEDIUM → HIGH

---

🚀 Quick Start

# Install
pip install -r requirements.txt

# Run Web UI
streamlit run scripts/streamlit_app.py

Models Required:

Model	Path	Purpose
YOLOv8 Detection	models/yolov8_leaf.pt	Leaf detection
MobileNetV2	models/mobilenetv2_disease.keras	Disease classification
YOLOv8 Segmentation	models/yolov8_seg.pt	Severity estimation ⭐

---

🎚️ Settings Guide

Setting	Value	Use Case
YOLO Confidence	0.35	General use ✅
Segmentation Toggle	ON	More accurate severity

---

👨‍💻 Author

Anuj Tripathi | EDP Project

---