Scientists yet again made a brilliant discovery on the surface of Mars, all thanks to the developing new technology AI.
It seems that between the time period of March 2010 and May 2012, a meteor broke into pieces and slammed onto the surface of the planet. This created a few small craters, which were around 4 meters in diameter.
It was very difficult to spot them given their relatively small size. However, this time, AI came to the rescue. This is a huge advancement indeed since now using technology, time shall be saved and it shall lead to more discoveries.
The issue with Context Cameras
Table of Contents
Generally, the brilliant spend many hours every day studying the images captured by NASA’s Mars Reconnaissance Orbiter (MRO). They look closely for evolving surface phenomena like avalanches, sand dunes, and dust devils. Scientists till now only relied on the MRO data and found about 1000 craters on the surface. Those craters are first detected using the Context Camera. The Context Camera takes low-resolution images and also it covers hundreds of miles at a time. So, the problem here is that the individual craters cannot be seen in the images, only the impact is clear.
The power of HiRISE
Now, here is where the HiRISE or High-Resolution Imaging Science Experiment enters. The images captured by the HiRISE are so powerful that you can view the tiniest details like the tracks left by the Curiosity Mars Rover.
Scanning a Context Camera Image requires a lot of patience. It takes atleast 40 minutes to carefully observe each and every detail in a single image.
Why did they create an AI tool?
To save the much needed energy and time, the researchers of JPL invented a tool. It is an automated fresh impact crater classifier coined COSMIC (Capturing Onboard Summarization to Monitor Image Change).
Researchers trained the tool by feeding it 6830 Context Camera images. Also, images with no fresh impacts have also been given to the classifier, to let it know what not to look for.
Once trained, the classifier was deployed on the Context Camera’s entire repository of 112,000 images. Running on a supercomputer cluster at the JPL, the AI tool completed the task in around five minutes. It is no ordinary feat since the process takes 40 minutes for a human.
Gary Doran a JPL computer scientist said:
“It wouldn’t be possible to process over 112,000 images in a reasonable amount of time without distributing the work across many computers.”
He informed the possible solution may be to split the problem into smaller parts and they can be solved in parallel.
Though the results might be very accurate with such high computing power, it still needs to be verified by a human.
Another JPL computer scientist Kiri Wagstaff informed that AI cannot do the kind of analysis that a scientist can.
“But tools like this new algorithm can be their assistants. This paves the way for an exciting symbiosis of human and AI ‘investigators’ working together to accelerate scientific discovery.”