Deriving planetary surface composition from orbiting observations from spacecraft


NASA has sent many robotic spacecraft to perform remote-sensing observations of surface composition from orbit. This includes gamma-ray observations of planetary surfaces, which provide position-dependent energy spectra whose shape is due to the sum of components from different elements. We seek to develop a machine learning (ML) approach to isolate the element-dependent contributions to the measurements as a function of surface location. We will use the Lunar Prospector Gamma-Ray Spectrometer dataset to train a model using the moon as the training dataset. The first objective is to identify the best ML model approach and quantify its accuracy as this data, and the Moon, are well understood. We will then use domain adaptation to extend the approach to other planetary objects that are less well known, enabling new discoveries.


Total project length: 175/350 hours.

Task ideas

Expected results


Python and relevant past experience in Machine Learning.


Please DO NOT contact mentors directly by email. Instead, please email with Project Title and include your CV and test results. The mentors will then get in touch with you.

Corresponding Project

Participating Organizations