This project provides a comprehensive guide to predicting burned areas on Landsat 5 images using machine learning in Python. It covers all the steps involved in the process, from image acquisition and data preparation to model training, evaluation, and prediction.
To run the notebook, you will need to install the following dependencies:
- pystac
- pystac_client
- planetory_computer
- rasterio
- geopands
- numpy
- scikit-learn
- matplotlib
- seaborn
You can install these dependencies using pip: pip install pystac pystac_client planetary_computer rasterio geopandas numpy scikit-learn matplotlib seaborn
To use the notebook, simply open it in Jupyter Notebook or JupyterLab and run the cells in order. The notebook includes detailed comments and explanations for each step of the process.
Both the Landsat 5 image and the trainining data is available in the data folder.
The notebook produces a random forest classifier that can predict burned areas on Landsat 5 images with an accuracy of over 85%. The notebook also includes visualizations of the data and the results.
If you find any issues with the notebook or would like to contribute to the project, please submit a pull request or open an issue on GitHub.
This project is released under the MIT License. See the LICENSE file for more information.