Project Case Study

Every Crater a Tale: Machine Learning Reveals the, Until Now, Hidden History of the Moon

CSIRO, Pawsey Supercomputing Centre, The Australian Space Data Analysis Facility, Space Science and Technology Centre.

A joint project between Perth’s Pawsey Supercomputing Centre, Curtin University and ASDAF is using crater detection algorithms to detect and analyse the billions of craters scarring the surface of the Moon to reveal its geological history.

The crater detection algorithm in question – originally designed by research engineers at CSIRO and scientists at Curtin University to analyse images from Mars – has been adapted and repurposed to fulfil the same role for the Moon. This effort has resulted in a revised algorithm that is tailor-fit to the specific requirements of analysing the Moon’s images and has produced the world’s largest database of crater size and location.

In the future, the algorithm would, ideally, be adapted to extract more information from the database, such as the conditions of the asteroids and the age of the craters. However, this database does not just open up a world of possibilities for future investigations on the Moon, the lessons learnt in this project could be applied to our world. A database and analysis of Earth’s topography could be used in environmental conservation, agriculture, predictions of natural disasters, and more.

The joint team – consisting of Kosta Servis (Pawsey/ ASDAF), Dr Anthony Lagain (Curtin University), Professor Gretchen Benedix (Curtin University) and John Fairweather (PhD Candidate, Curtin University) – is expecting to complete the project in March 2022. Currently, the algorithm is running at Pawsey, where it is analysing the images and compiling data to be compared to a baseline data set of craters that was painstakingly counted and analysed by the team.