How machine learning is unlocking five decades of satellite imagery from space

NASA has been collecting global data via the Landsat mission for over 50 years. By applying modern machine learning and deep learning models to this massive archive, researchers are now mapping environmental changes and human impact with unprecedented accuracy and efficiency.

The Landsat mission has provided a continuous stream of Earth observation data for more than half a century. Recently, scientists have begun leveraging artificial intelligence to process this vast repository, moving beyond manual analysis to automated land cover classification and landscape change detection.

Current research focuses on refining deep learning architectures to extract more nuanced insights from the imagery. This evolution in data processing allows for a more comprehensive understanding of how human activity shapes the environment, providing a clearer picture of the planet's history and its potential future trajectories.

Source: From Orbit to A.I. - Harnessing Machine Learning with Landsat Data