CropScan IR: Smart Crop Monitoring System Using Thermal Imaging

CropScan is an experimental AI-based agricultural monitoring tool that leverages thermal imaging and machine learning to analyze crop health conditions including:

• 🌿 Pest infestation and disease detection
• 💧 Irrigation need estimation
• 🌫️ Humidity-related stress prediction
• 🌾 Harvest time prediction

Built with a MERN stack (MongoDB, Express.js, React.js, Node.js) and integrated with a custom CNN + Random Forest ML model, this system offers a unified pipeline from image acquisition to intelligent decision-making.

---

🚀 Features

• 📷 Accepts thermal images as input
• 🧠 Extracts statistical, gradient, and histogram-based temperature features
• 🔍 Uses CNN for image-based pattern recognition
• 📊 Classifies pest, irrigation, and humidity conditions via multi-output random forest
• 🌐 Integrated API layer using Flask and Node.js backend

---

🛠️ Tech Stack

• Frontend: React.js
• Backend: Node.js + Express
• Database: MongoDB (Atlas)
• ML Inference: Python Flask (CNN + RandomForestClassifier)
• Libraries: OpenCV, TensorFlow/Keras, NumPy, scikit-learn

---

📦 Installation

1. Clone the Repository

   git clone https://github.com/sourav-625/CropScan.git
   cd CropScan

2. Install backend dependencies

   cd backend
   npm install

3. Set up a Python virtual environment

   cd ../model
   python -m venv venv
   source venv/bin/activate

4. Install dependencies

   pip install -r requirements.txt

5. Install frontend dependencies

   cd ../
   npm install

Run The Application

To Run the application first run the model.py and then start the react app (NOTICE: You must have python 3.7 or higher installed in your system)

python ./model/model.py
npm start

Now you can view the react app opened in your browser on http://127.0.0.1:3000/

📬 Contributing

We welcome contributions from researchers, engineers, and agritech developers. Please feel free to open issues or submit a pull request after forking this repository and committing your changes on a separate branch.

⚠️ Disclaimer This project is currently in its experimental and theoretical stage. The ML model and its predictions are based on synthetically generated or simulated data. It has not been validated in real-world agricultural environments. Before real deployment, it must be rigorously tested, verified by agricultural scientists, and certified by experts in crop science and thermal imaging. It is not recommended for critical decision-making in production-level agriculture without professional review.