Oxia Planum, Mars, classified using the NOAH-H deep-learning terrain classification system

• DOI: 10.1080/17445647.2022.2112777
• Publication Date: 2022-08-18T19:44:06Z
• Authors: Alexander M. Barrett; Jack Wright; Elena Favaro; Peter Fawdon; Matthew R. Balme; Mark J. Woods; Spyros Karachalios; Eleni Bohacek; Elliot Sefton-Nash; Luc Joudrier
• Keywords: Planum temporale

We present a map of Oxia Planum, Mars, the landing site for ExoMars Rover. This shows surface texture and aeolian bedform distribution, classified using deep learning (DL) system. A hierarchical classification scheme was developed, categorising textures observed at site. then used to train DL network, 'Novelty or Anomaly Hunter – HiRISE' (NOAH-H). The applied across wider area than could have been mapped manually. result showed strong agreement with human-mapped areas reserved validation. resulting product is presented in two ways, representing principle levels scheme. 'Descriptive classes' are purely textural nature, making them compatible machine approach. These combined into 'interpretive groups', broader thematic classes, which provide an interpretation landscape. step allows more intuitive analysis results by human users.