In-orbit Demonstration Of A Re-trainable Machine Learning Payload For Processing Optical Imagery

This article discusses the use of machine learning enabled satellite networks, which are self-contained software modules that are encapsulated in a virtualised software container. ML payloads offer a simple pathway to upgrade, correct or enhance satellite capabilities in a relatively risk-free way. The article also describes the methodology and experimental setup used to benchmark the models in different edge devices, as well as the results of the payload execution in orbit.