Scientists use AI technology to obtain clearer images of lunar craters

According to foreign media reports, the moon is a cold and dry desert. Unlike Earth, there is no protective atmosphere around the moon, and water that existed during its formation evaporated for a long time under the influence of solar radiation and escaped into space. However, the circular mountains and depressions in polar regions give people reason to hope that water resources are limited.

Scientists from MPS, Oxford University, and NASA Ames Research Center have now conducted closer observations of some of these areas.

"Near the South and North Poles of the Moon, incoming sunlight enters craters and depressions at very shallow angles, never reaching some of their surfaces," said MPS scientist Valentin Bickel, the lead author of this study. On this eternal night, temperatures in some places are so low that frozen water is expected to last for millions of years - possibly due to impacts from comets or asteroids, gas released from volcanic eruptions, or the interaction between the surface and the solar wind. In recent years, measurements of neutron flux and infrared radiation by space probes have shown the presence of water in these areas. In the end, NASA's Lunar Crater Observation and Sensing Satellite (LCROSS) provided direct evidence: 12 years ago, the probe fired a shell at the gloomy Cabeus crater in Antarctica. Subsequent analysis showed that the dust and mist emitted into space contained a large amount of water.

However, permanently shaded areas are not only hot topics in scientific research. If humans were to stay on the moon for a long period of time, naturally occurring water would be a valuable resource - and shadowed craters and depressions would be an important destination. For example, NASA's unmanned VIPER probe will explore the Antarctic region and enter these craters in 2023. In order to obtain accurate terrain and geological images in advance, such as for mission planning purposes, images from space probes are essential. NASA's Lunar Reconnaissance Orbiter (LRO) has been providing such images since 2009.

However, capturing images in the deep darkness of permanently shaded areas is extremely difficult, as the only light source is scattered light. The co-author of this study, Ben Moseley from the University of Oxford, pointed out that "due to the spacecraft being in motion, LRO images will be completely blurred after prolonged exposure." In shorter exposure times, spatial resolution is much better. However, due to insufficient lighting, these images are mainly noisy, making it difficult to distinguish real geological features.

To address this issue, researchers have developed a machine learning algorithm called HORUS (Hyper effective nOise Removal U-net Software), which can "clear" these noisy images. It is understood that it used over 70000 calibration images taken by LRO on the dark side of the moon, as well as information on camera temperature and spacecraft orbit, to distinguish which structures in the images are artificial and which are real. Through this method, researchers can achieve a resolution of approximately 1-2 meters per pixel, which is 5 to 10 times higher than the resolution of all previously available images.

By using this method, researchers have now reassessed images of 17 shaded areas in the South Pole region of the Moon, with an area ranging from 0.18 to 54 square kilometers. In the resulting images, only a few meters wide small geological structure can be more clearly identified than before. These structures include round stones or very small craters, which can be seen everywhere on the surface of the moon. Due to the lack of an atmosphere on the moon, very small meteorites constantly fall on the surface of the moon and form these miniature craters.

"With the help of the new HORUS images, it is now possible to have a better understanding of the geological conditions of the lunar shadow areas than before," Moseley said. The number and shape of small craters provide information about the age and composition of the Earth's surface. It can also more easily identify potential obstacles and dangers for wanderers or astronauts. Researchers have discovered a very bright micro crater in a studied crater in Leibnitz Plateau. Bickel said, "Its relatively bright color may indicate that this crater is relatively young." Researchers suggest that because such a newly left scar can easily penetrate deeper, this location may be an interesting target for future missions.

The new image does not provide evidence of frozen water on the lunar surface. Bickel speculated that "some of our target areas may be slightly too hot." The water on the lunar surface is likely not present at all, but mixed with weathering and dust or hidden underground.

To address this and other issues, the next step for researchers is to use HORUS to study as many shaded areas as possible. "In the current publications, we want to demonstrate what our algorithm can do. Now we want to apply it as comprehensively as possible," Bickel said.

Prev: Scientists have created a micro robot for drug delivery driven by bubbles controlled by ultrasound