https://github.com/mjahmadee/eurosat_deeplearning

Analysis and classification of satellite images. A novel dataset and deep learning benchmark for land use and land cover classification (Eurosat)
https://github.com/mjahmadee/eurosat_deeplearning

image-classification satellite-data satellite-imagery satellite-images transfer-learning

Last synced: 2 months ago
JSON representation

Analysis and classification of satellite images. A novel dataset and deep learning benchmark for land use and land cover classification (Eurosat)

• Host: GitHub
• URL: https://github.com/mjahmadee/eurosat_deeplearning
• Owner: MJAHMADEE
• License: mit
• Created: 2023-07-13T11:21:17.000Z (about 2 years ago)
• Default Branch: main
• Last Pushed: 2024-03-16T12:58:47.000Z (over 1 year ago)
• Last Synced: 2025-02-28T20:45:15.358Z (7 months ago)
• Topics: image-classification, satellite-data, satellite-imagery, satellite-images, transfer-learning
• Language: Jupyter Notebook
• Homepage:
• Size: 10.1 MB
• Stars: 4
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Eurosat Deep Learning for Land Use and Land Cover Classification 🌍🛰️

![Python](https://img.shields.io/badge/Python-3.x-blue.svg)

![PyTorch](https://img.shields.io/badge/PyTorch-1.x-orange.svg)

![Deep Learning](https://img.shields.io/badge/Deep%20Learning-CNN-green.svg)

This project aims at classifying land use and land cover from the Eurosat dataset using Deep Learning techniques. The dataset comprises satellite images from the Sentinel-2 mission, which are used to train a Convolutional Neural Network (CNN) for image classification.

## Features 🌟

- Utilizes the Eurosat dataset for training and testing the model.

- Employs the VGG-16 architecture with modifications to adapt to the number of classes and input channels.

- Provides detailed data loading and preprocessing to handle multi-band satellite images.

- Offers insights into the model's performance through accuracy, precision, recall, F1-score, and a confusion matrix.

- Visualizes training progress, class distributions, and predictions.

## Setup and Installation 🛠️

1. Clone the repository.

2. Install the necessary Python packages listed in `requirements.txt`.

3. Download the Eurosat dataset and prepare it according to the instructions provided.

## Data 📁

The Eurosat dataset contains labeled satellite images covering 10 different classes of land use and land cover. Images are in TIFF format with multiple spectral bands.

## Model Training and Testing 🚀

- The model is trained using a pre-processed subset of the Eurosat dataset.

- Training includes several epochs with batch processing, validation checks, and performance logging.

- Testing is performed to evaluate the model's accuracy and generalization on unseen data.

## Results and Evaluation 📊

- Performance metrics are calculated for the test dataset to evaluate model accuracy.

- A confusion matrix is generated to understand the classification performance across different classes.

- Sample images with predictions are displayed to visualize the model's capabilities.

## Contributing 🤝

We welcome contributions to improve the project. Feel free to fork the repository, make your changes, and submit a pull request.

## License 📜

The project is licensed under the MIT License - see the LICENSE file for more details.

## Acknowledgements 🙌

- The Eurosat dataset providers for creating and distributing such a valuable resource for satellite image analysis.

- The PyTorch team for providing an excellent deep learning framework.

For more information and updates, visit the [GitHub repository](https://github.com/MJAHMADEE/Eurosat_DeepLearning/).