Unsupervised and supervised learning for satellite image classification
This repo contains real-life examples of machine learning applications (unsupervised and supervised learning) using satellite raster data.
The examples are based on unsupervised and supervised classification to investigate different landcovers.
The classifications are tested with Landsat 4 TM, Landsat 8 OLI, Landsat 7 ETM+, Sentinel 2 MSI, and Sentinel 3 OLCI optical satellite data. All data used for testing is publically available under open source license. For more details look here https://scihub.copernicus.eu/
Any of these classfications in this repo can be applied with any optical satellite data, from space-borne or air-borne sensors.
To run any of these jupyter file, user needs:
Anacond installed with Python 3.6 or above
Rasterio python package for working with satellite raster data
To install rasterio see here: https://rasterio.readthedocs.io/en/latest/
All machine learning applications are based on SciKit-Learn API https://scikit-learn.org/stable/modules/classes.html