An automatic classification method for mapping Martian landforms

Abstract

The physiographic map can visualize spatial relations between different landforms, thus providing insights into geologic processes that shaped the present-day Martian landscape. The physiographic map of Mars surface is usually made through image interpretation, which is always labor-intensive and highly depends on the expert knowledge. In this paper, we propose an efficient and automatic classification method for characterization of landforms on Mars by using the Mars Orbiter Laser Altimeter (MOLA) digital elevation data. The proposed method was tested on a region where China's Mars probe Tianwen-1 landed. The study area covers the Nepenthes Mensae, Amenthes Planum, northern Terra Cimmeria, northern Hesperia Planum and southern Utopia Planitia region, having a size of 2250km×2750km centered at 117°E, 6°N. The obtained results confirm the effectiveness of the proposed method in describing different topographic characteristics of the Martian landforms. Note that the proposed method is completely data-driven, which can provide a rapid mapping result in large geographical regions, especially from a global perspective to reveal the Martian landform information.