Mmabiaa/Illegal-Mining-site-Detection-Model

Illegal mining (galamsey) is a major environmental issue in Ghana, causing deforestation, water pollution, and land degradation. This project uses Convolutional Neural Networks (CNNs) and satellite imagery to automatically detect illegal mining areas. # Detection and Monitoring of Illegal Mining Activities Using Machine Learning and Remote Sensing Data ## 📌 Project Overview Illegal mining (galamsey) is a major environmental issue in Ghana, causing deforestation, water pollution, and land degradation. This project uses Convolutional Neural Networks (CNNs) and satellite imagery to automatically detect illegal mining areas. --- ## 🎯 Objectives * Detect illegal mining areas from satellite images * Differentiate between mining and non-mining regions * Demonstrate the use of Artificial Intelligence in environmental monitoring --- ## 🧠 Methodology ### 1. Data Collection We used a publicly available dataset from Kaggle containing satellite images of: * Illegal Mining areas (Garimpo) * Natural land (Rios_Floresta) ### 2. Data Preprocessing * Images resized to **150x150 pixels** * Pixel values normalized (0–255 → 0–1) * Dataset split into: * 80% Training * 20% Validation ### 3. Model Used A **Convolutional Neural Network (CNN)** was implemented to classify images. #### Model Architecture: * Input Layer * Conv2D (32 filters) + MaxPooling * Global Average Pooling (reduces overfitting) * Dense Layer (64 neurons) * Dropout (0.5) * Output Layer (Sigmoid) --- ## 🛠️ Technologies Used * Python * TensorFlow / Keras * NumPy * Matplotlib * Jupyter Notebook --- ## 📂 Project Structure ``` illegal-mining-project/ │ ├── dataset/ │ ├── illegal/ │ └── legal/ │ ├── illegal_mining.ipynb ├── test.jpg └── README.md ``` ## Results * Training Accuracy: ~99% * Validation Accuracy: ~95–97% * Model successfully distinguishes illegal mining areas ## Limitations - Limited dataset size - Slight overfitting observed - Performance depends on image quality - No real-time detection yet --- ## 🚀 How to Run the Project ### Step 1: Install Dependencies ```bash pip install tensorflow numpy matplotlib notebook ``` ### Step 2: Start Jupyter Notebook ```bash python -m notebook ``` ### Step 3: Open Notebook * Open `illegal_min …