Proposer: University of Kent and ACRI-ST

Thanks to rapid advances in imaging and communication technology, space science missions can acquire faster and more abundant streams of data than ever before. This particularly applies to the remote-sensing data retrieved by satellites in orbit around planetary bodies such as the Moon and Mars. Machine Learning (ML) and other computer vision techniques provide the opportunity to analyse this data in shorter times and with higher accuracies than is achievable for humans. However, while there is an increase in the use of ML in planetary science, it is not accelerating at the same pace as seen in other fields such as geophysics or astronomy (Azari et al. 2020).

It is the goal of this PhD project to develop ML tools which can automatically detect planetary surface features, in order to fulfil the need of exploiting the vast volumes of space data currently available. This project is a collaboration between the University of Kent, ACRI-ST (an SME of the space sector), and the Observatoire de la Côte d’Azur.

Acknowledgements: This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 101004214.

References:

• Azari, A.R.; Biersteker, J.B.; Dewey, R.M.; Doran, G.; Forsberg, E.J.; Harris, C.D.; Kerner, H.R.; Skinner, K.A.; Smith, A.W.; Amini, R.; Cambioni, S. Integrating machine learning for planetary science: Perspectives for the next decade, 2020, arXiv:2007.15129.

Martian Pit Shadow extractor (MAPS) is an automated tool for detecting the shadows cast by the Sun into Martian pits in order to calculate the pits’ apparent depth. Martian pits are circular-to-elliptical depressions on the surface of Mars, which are most likely caused by gravitational collapse into a sub-surface void. MAPS employs an adaptation of the Benjamini-Hochberg procedure to extract the shadow from a cropped red-band image taken of a Martian pit by the Mars Reconnaissance Orbiter (MRO) HiRISE camera. However, when information about the surface surrounding the pit is unavailable, MAPS uses K-Means clustering for shadow extraction instead. By measuring the width of the extracted shadow, MAPS calculates the pit’s apparent depth, which is the depth at the edge of the shadow (Wyrick et al. 2004). Martian pits are candidate entrances to networks of evacuated lava tubes (Cushing et al. 2007) and the depth of the pit is a factor that will influence the volume of intact lava tubes underneath the surface (Sauro et al. 2020). The propensity for ice caves on Mars increases with the thickness of the cavity’s ceiling (Williams et al. 2010), which itself will be larger for deeper pits. MAPS will soon be adapted to be able to analyse pits found in data taken from other space missions as well as data of other planetary bodies.

References:

• Wyrick, D.; Ferrill, D.A.; Morris, A.P.; Colton, S.L.; Sims, D.W. Distribution, morphology, and origins of Martian pit crater chains, Journal of Geophysical Research, 2004, 109, E06005, doi:10.1029/2004JE002240.
• Cushing, G.E.; Titus, T.N.; Wynne, J.J.; Christensen, P.R. THEMIS observes possible cave skylights on Mars, Geophysical Research Letters, 2007, 34, L17201, doi:10.1029/2007GL030709.
• Sauro, F.; Pozzobon, R.; Massironi, M.; De Berardinis, P.; Santagata, T.; De Waele, J. Lava tubes on Earth, Moon and Mars: A review on their size and morphology revealed by comparative planetology, Earth-Science Reviews, 2020, Vol. 209, 103288, ISSN 0012-8252, doi.org/10.1016/j.earscirev.2020.103288.
• Williams, K.E.; McKay, C.P.; Toon, O.B.; Head, J.W. Do ice caves exist on Mars?, Icarus, 2010, Vol. 209, Issue 2, pp 358-368, ISSN 0019-1035, doi.org/10.1016/j.icarus.2010.03.039.

Work has also begun on a project which will develop a tool for detecting channels and gullies on the surface of Mars, which can provide information about the abundance of past and present surface water content. This work involves converting scene classifications of these features in MRO CTX camera images made as part of the DoMars 16k survey (Wilhelm et al. 2020) into polygon labels and training a model classify on a pixel-level and estimate their global surface areas.

References:

• Wilhelm, T.; Geis, M.; Püttschneider, J.; Sievernich, T.; Weber, T.; Wohlfarth, K.; Wöhler, C. DoMars16k: A Diverse Dataset for Weakly Supervised Geomorphologic Analysis on Mars, Remote Sensing. 2020, 12, 3981. doi.org/10.3390/rs12233981