Experts of laboratory of jet movement NASA have launched a new citizen science project AI4Mars, in which anyone can help in the learning algorithm of automatic recognition of terrain for Mars Rovers “Curiosity” and “Perseverance”. You need to view pictures taken by the Rovers, and mark different types of terrain, reported on the website of the jet propulsion Laboratory of NASA.
The main danger for Rovers is the landscape of Mars. Numerous sand dunes caused the death of the Mars Rover “Spirit” in 2011 and repeatedly interfere with the normal operation of the Rovers “opportunity” and “Curiosity”, and sharp stones along the way that can damage the wheels of the Rovers. Planning a one-day trip “Curiosity” team of specialists takes several hours, since it is necessary to identify interesting sites for geological work, avoid dangerous areas and to make the end of the route the Rover could reach the Ground.
To facilitate the work of planners from the team “Curiosity”, the specialists from laboratory of jet movement NASA have developed a classification algorithm location SPOC (Soil Property and Object Classification), in which the Rover can distinguish between dense soil, high rocks, smooth rock and sand dunes. In order to improve the algorithm, we need to continue his training, but the difficulty lies in the necessity of creating a large database of images that are already marked with all types of terrain.
A group of researchers headed It Hiro (Hiro Ono) from the jet propulsion Laboratory, NASA launched a project AI4Mars on the basis of the world’s largest citizen science portal Zooniverse. Project database currently contains over eight thousand photographs taken “Curiosity”, in the near future expected download images taken with the Rovers “Spirit” and “opportunity”.
To participate in the project can be anyone with access to the Internet. To get started you need to register with the website and to read the rules. The aim of this work is the marking of different types of terrain in the images, then the resulting sample will be tested and used to further train the algorithm, which engineers want to apply for future work with the Rover “Perseverance”, which should land on Mars in February 2021.
Earlier we talked about the project AstroQuest, in which you can participate in the data processing the largest spectroscopic galaxy survey WAVES, and also about Hubble Asteroid Hunter, which you need to look for the tracks of asteroids in the images of the space telescope “Hubble”.