🌍 Satellite Image Segmentation for Deforestation Detection

📌 Overview

This project focuses on deforestation detection using satellite image segmentation techniques. It employs K-Means, Mean Shift, and Hierarchical Clustering to segment images and identify deforested areas. The results contribute to environmental conservation by providing insights for decision-makers and researchers.

🏆 Key Features

📡 Satellite Image Segmentation to monitor forest loss 🏗 Clustering Algorithms (K-Means, Mean Shift, Hierarchical) 📊 Performance Evaluation using Intersection over Union (IoU) 🌲 Real-World Dataset (2013-2024, Chandrapur region) 🛠 Python & OpenCV-based Implementation

🛠 Methodology

The project applies three clustering techniques:

1️⃣ K-Means Clustering

Groups pixels based on color similarity Identifies deforested areas using pixel intensity analysis

2️⃣ Mean Shift Clustering

Detects spatially cohesive clusters Assumes lower green intensity represents deforestation

3️⃣ Hierarchical Clustering

Constructs a tree-like structure for segmentation Effective for small-scale deforestation analysis

📊 Results

K-Means: Detected 183.5 sq. meters deforested in 2023

Mean Shift: Identified 258.25 sq. meters

Hierarchical Clustering: Found minor deforestation changes

🚀 Installation & Usage

Prerequisites Python 3.x Jupyter Notebook OpenCV, NumPy, Matplotlib, Scikit-Learn

📜 Conclusion

This project provides an automated approach for detecting deforestation using unsupervised clustering techniques. The results help track forest loss over time and contribute to environmental conservation efforts.