The American Mars rover Curiosity has been crossing the Mars landscapes in the Gale Crater since 2012. Its successor Perseverance (also known as the Mars 2020 Rover) and China's Tianmen-1 Rover Mission are on their way to land on Mars in February and May 2021 respectively, and the European Rosalind Franklin Rover will follow in 2023. They will focus on the search for life and water; Perseverance will even attempt to collect samples to be brought to Earth. In addition to these main objectives, the missions are dedicated to a variety of other scientific tasks. One of these is the search for rocks that allow the detection of impact structures on Mars.
The project Mars-DL (Planetary Scientific Target Detection via Deep Learning), funded by the FFG ASAP program, in which VRVis, Joanneum Research, Naturhistorisches Museum Wien, SLR Engineering and the University of Vienna participated, dealt with important aspects of such a search: Where and how can images from Mars rovers be used to detect such impact craters? What are the criteria for such a search? Can we use novel artificial intelligence techniques to support such a search? Can we use powerful computing power to train such systems by letting computers literally teach themselves to deliver new research results for these highly complex reconnaissance missions?
One of the presentations will be held by Christoph Traxler, project manager of VRVis. His talk "Visualization and Simulation for Artificial Intelligence (AI) Training in Space Research" will explore the question whether deep learning (DL) can be used for autonomous target selection of future Mars rover missions to assist planetary scientists in pre-selecting potentially interesting regions, thereby increasing scientific discoveries and accelerating strategic decision-making.
For further information