🚀 Getting Started

Prerequisites

Step 1 — Start the Backend

cd "Smart Crop, Fertizer and Disease Prediction System/agri-tools/backend"
mvn spring-boot:run

✅ You should see:

Crop dataset loaded: 100 records
Tomcat started on port 9090
Started AgriToolsApplication in ~3 seconds

Step 2 — Open the Frontend

Double-click agri-tools/frontend/index.html in File Explorer.

⚠️ Keep the backend terminal open while using the app. The frontend calls http://localhost:9090/api.

Step 3 — Verify Backend

Visit this in your browser to confirm the API is live:

http://localhost:9090/api/fertilizer/crops

Expected:

{"crops":["rice","wheat","maize","cotton","mango","banana","sugarcane"]}

🌐 API Reference

🌾 Crop Prediction

POST http://localhost:9090/api/crop/predict
Content-Type: application/json

Request:

{
  "nitrogen": 83,
  "phosphorus": 33,
  "potassium": 36,
  "ph": 6.21,
  "rainfall": 68.7,
  "region": "North"
}

Response:

{
  "recommendedCrop": "cotton",
  "region": "North",
  "message": "Prediction based on k-NN (k=5) over 100 records."
}

🧪 Fertilizer Recommendation

POST http://localhost:9090/api/fertilizer/recommend
Content-Type: application/json

Request:

{
  "nitrogen": 60,
  "phosphorus": 50,
  "potassium": 30,
  "crop": "rice"
}

Response:

{
  "crop": "Rice",
  "optimalNpk": { "N": 90, "P": 40, "K": 40 },
  "currentNpk": { "N": 60.0, "P": 50.0, "K": 30.0 },
  "adjustments": [
    "Increase Nitrogen by 30 units (current: 60, target: 90).",
    "Reduce Phosphorus by 10 units..."
  ],
  "status": "needs_adjustment"
}

🔬 Disease Detection

POST http://localhost:9090/api/disease/detect
Content-Type: multipart/form-data

image: <file>   (JPG / PNG / WEBP — max 10 MB)

Response:

{
  "disease": "Bacterial Spot",
  "description": "A common disease caused by Xanthomonas bacteria...",
  "treatmentSteps": [
    "Use certified pathogen-free seed and disease-free transplants.",
    "Rotate crops to break the bacterial cycle.",
    "Apply copper-containing bactericides regularly."
  ],
  "severity": "moderate"
}

🧪 Sample Test Data

🌾 Crop Recommendation — 10 Examples

🧪 Fertilizer Guidance — 10 Examples

🔬 Disease Detection — 10 Examples

Rename any leaf image to include these keywords before uploading:

💡 The classifier checks the filename first. Rename any image with the keyword above to trigger that specific result. Replace with a real CNN model by updating classifyStub() in DiseaseDetectionService.java.

🧠 ML Architecture

Crop Prediction — k-NN (k=5)

• Loads recommended_crops_dataset.csv (100 records) at startup
• Features: Nitrogen, Phosphorus, Potassium, pH, Rainfall, Region (encoded 0–4)
• Applies min-max normalisation per feature before computing Euclidean distances
• Returns the majority-vote label among the 5 nearest neighbours

Fertilizer Recommendation — Rule-Based

• Compares user N-P-K against hardcoded optimal values per crop
• Outputs precise delta adjustments (increase/decrease by X units)
• Supported crops: rice, wheat, maize, cotton, mango, banana, sugarcane

Disease Detection — Stub + Knowledge Base

• 4 diseases in knowledge base: Bacterial Spot, Leaf Blight, Powdery Mildew, Healthy
• Each has description, severity, and step-by-step treatment
• Current classifier: keyword-based stub (replace with CNN for production)

🔄 Improvements over Original Python Project

📄 License

MIT License — free to use, modify, and distribute.

👤 Author

Tamilarasan Palanisamy Full Stack Developer — Java Spring Boot + HTML/CSS/JS

AI -powered monitoring of crop health ,soil conditions and pest risk using multispectral/hyperspectral imaging and sensor data using Python → AI/ML model, backend, sensor data processing HTML, CSS, JavaScript → Frontend dashboard UI Domain name : Artificial intelligence machine learning