Some regions close to the moon's poles never receive direct sunlight and remain permanently under shadow. According to recent research, these so-called persistently shadowed regions (PSRs) may have rich ice reservoirs that might allow future visitors create fuel and other resources while also providing information about the early solar system. However, these regions are difficult to examine since they are difficult to image from satellites orbiting the moon. Static-like camera noise and quantum phenomena frequently overwhelm the few photons that PSRs do reflect.
To see these dark zones and cut through the noise, researchers have now developed a deep learning method. Valentin Bickel, a planetary scientist at the Max Planck Institute of Solar System Research in Germany and the lead author of a Nature Communications study testing the new algorithm, says, "Our images enable scientists to identify geologic features, such as craters and boulders... as small as three meters across for the first time a 5 to 10 fold increase in resolution compared to previous efforts."
To train their algorithm to identify and eliminate camera noise, the researchers used over 70,000 photos of entirely black lunar regions with no light signal along with information regarding the camera's temperature and orbital position. Subsequently, they addressed residual noise, including quantum effects on moving photons. This algorithm stage was trained using millions of lunar shots taken in the sun and simulated, shadowed copies of the same images.
According to Ignacio Lopez-Francos, an engineer at the NASA Ames Research Center and co-author of the paper, the lack of sunlight PSR photos necessitated the creation of this artificial shadow. Digital camera photography in low light also employs a similar method.
According to computer scientist Chongyi Li, who was not involved in the study but employs similar techniques to improve underwater photos at Singapore's Nanyang Technological University, "it is an interesting application of machine-learning technology, and the noise model seems realistic and useful for this real case."
The size and distribution of craters and boulders in a number of PSRs that NASA's Artemis moon program may investigate were examined by the researchers using their technique. Plotting a possible path for a rover through a PSR on the moon's Leibnitz plateau, avoiding obstructions and slopes steeper than 10 degrees, they also assessed the probable origins of a few rocks.
Jose Hurtado, a geologist at the University of Texas at El Paso who was not involved in the study, says, "There is a lot of curiosity in the poles—not just from the human exploration aspect but also the topography of the ground surface." According to him, the terrain may be distorted by ice that is either scattered throughout the lunar soil or held in more concentrated strata. Consequently, this type of image processing.
@ Bhautik Thummar