25 December 2020
In Nature Communications an international team of lunar scientists published an article about how Chang'e 1 and Chang'e 2 data and stratigraphic information was used for automatic crater detection and age estimation. The researchers applied machine-based transfer learning where one data model serves as the starting point for "training" a neural network for solving incremental advanced problems. Through this approach the scientists identified 109,956 new craters which is more than a dozen times greater than the initial number of recognised craters. Additionally, they also estimated the age of 18,996 newly detected craters larger than 8 km in diameter. As a result, a new database for lunar craters in the mid- and low-latitude regions of the Moon could be established.
link to publication in Nature Communications