Several Martian rovers have already succumbed to the terrible sand dunes on Mars. To prevent this from happening to Curiosity or Perseverance in the future, NASA engineers invite you to help them classify photos of Martian soil so that rovers can better identify them and which ones to avoid.
NASA needs you! Engineers from the United States Aerospace Agency have launched a platform that aims to make Martian rovers, like Curiosity, able to better recognize the terrain on which they move. Indeed, over the years and generations, rovers have often finished their “road trip” mired in the sand on Mars. This was the case for Spirit in 2009 and, to a lesser extent, for Opportunity, which got stuck in the sand for six weeks. To improve the field analysis skills of Curiosity and future rovers, the engineers created the AI4Mars initiative. Completely open to the public, it proposes to categorize different surfaces of terrestrial and Martian soil, like what a photographic identification captcha would require.
The analysis data thus already ready and provided by benevolent volunteers will be used to enrich SPOC (for “Soil Property and Object Classification”). This artificial intelligence, which Curiosity benefits from and which will also be installed on Perseverance (whose launch is scheduled for July 20), allows rovers to classify the types of soil and terrain. Thanks to SPOC, rovers orient themselves better in space and more easily avoid danger zones like sand dunes. The AI4Mars platform includes more than 8,000 photos of the planet Mars taken by the Curiosity rover and which need to be characterized before being injected into SPOC. The AI4Mars tool allows you to form colored polygons in each photo to be associated with four types of terrain according to the grain size: sandy, with fine gravel, with pebbles (less than 30 cm thick) or with large rocks (more than 30 centimeters). Ideally, it’s best for the rover to move over small gravel or pebbles.