Sometime between March 2010 and May 2012, a meteor crossed the Martian sky, breaking into pieces and crashing into the surface of the Red Planet. The resulting craters are small, 13 feet in diameter. The smaller the features on the planet’s surface, the harder it will be to find them from the Mars orbiter. Now, for the first time, scientists have discovered the crater using artificial intelligence and machine learning.
Planetary scientists say the use of artificial intelligence in such sophisticated astronomical research is a milestone. Artificial intelligence researchers from NASA’s Jet Propulsion Laboratory have developed a machine learning tool to detect impact craters. The researchers hope the new AI will save time and increase discovery.
Scientists have found that a typical method of such craters is to spend hours a day studying images taken by the Mars Reconnaissance Orbiter and looking for surface movements. Scientists working on Mars have discovered more than 1,000 new craters in their 14-year orbit, relying on data from the Mars Reconnaissance Orbiter.
In the image above, only the explosive marks around the impact are prominent and individual craters cannot be seen. The next step will be to observe the area using a high-resolution imaging science experiment called HiRISE. The instrument is powerful enough to see details as fine as the footprints left by the Curiosity rover.
The researchers manually searched for surface phenomena using photographs, and scanning a contextual camera image took about 40 minutes. To speed up the process, the researchers created a tool called the Automatic Fresh Impact Crater Classifier. Training the classifier required researchers to provide them with 6,830 images before and after the event, including those previously identified by HiRISE. Humans take 40 minutes to analyze an image, while artificial intelligence tools take only five seconds. The researchers point out that although classifiers have all the computing power, they still need to be examined by humans.