🌍 Factory Emissions Visualizer

This project is a satellite-powered environmental monitoring tool that detects abnormal thermal emissions around industrial areas. Built using Google Earth Engine, Python, Streamlit, Folium, and ML-based anomaly detection, it identifies temperature hotspots, scores emission severity, and visualizes results on an interactive map.

---

🔥 Why This Project Is Useful

• Detects thermal anomalies caused by waste burning, leaks, illegal emissions, or overheated machinery
• Uses real satellite data — unbiased and independent of on-site sensors
• Ideal for government regulators, smart-city systems, ESG analysis, and industrial audits
• Fully automated pipeline: Fetch → Detect → Score → Visualize
• Lightweight, fast, and easy to extend

---

🛠️ Tech Stack

• Satellite Data: Google Earth Engine (Landsat-9 LST)
• ML Model: Isolation Forest (Scikit-Learn)
• Frontend/UI: Streamlit
• Geospatial Visualization: Folium + Matplotlib
• Backend Logic: Python
• Deployment Ready: Works locally or on cloud platforms (Streamlit Cloud, GCP, etc.)

Factory-Emission-Visualizer/
 ├── app.py
 ├── anomaly.py
 ├── fetchLST.py
 ├── folium_map.py
 └── key.json            # GEE service account credentials (DO NOT COMMIT)

---

📦 File-by-File Overview

1. app.py — Main Application (Streamlit UI)

• Takes user input (latitude & longitude)
• Fetches LST data from Google Earth Engine
• Runs anomaly detection and computes emission score
• Displays metrics (max/min/mean LST, anomaly count, emission score)
• Renders an interactive Folium-based heatmap

This is the main entry point of the project.

---

2. anomaly.py — ML-Based Temperature Anomaly Detector

• Downloads .npy temperature grid
• Removes invalid pixels
• Detects hotspots using Isolation Forest
• Generates a 0–100 emission severity score

The core intelligence/ML logic lives here.

---

3. fetchLST.py — Google Earth Engine Data Fetcher

• Authenticates using a service account
• Filters Landsat-9 LST images (cloud-free, last 30 days)
• Converts DN → Kelvin → °C
• Produces a downloadable .npy array

This module retrieves real satellite data.

---

4. folium_map.py — Geospatial Visualization Engine

• Uses Matplotlib to render temperature heatmap
• Highlights anomaly pixels
• Converts plot into base64 image
• Displays it as a Folium overlay with markers

This file generates the interactive visual output.

---

🚀 Running the Project

1. Install dependencies

Manually install the required libraries:

pip install streamlit folium scikit-learn numpy matplotlib google-earth-engine

2. Add your Earth Engine credentials

Place your key.json file in the project folder.

3. Run the Streamlit app

streamlit run app.py

4. Enter factory coordinates

Example:

Latitude: 20.95150000
Longitude: 85.21570000

Click Analyze Emissions to view:

• Heatmap overlay
• Detected anomaly pixels
• Emission severity score
• Summary statistics

---

📌 Future Improvements

• Time-series anomaly tracking
• Multi-factory comparison
• High-resolution LST (ECOSTRESS) support
• Automated pollution alerts
• Historical emission reporting dashboard

---

📄 License

This project is fully open-source and free to modify.