Tokyo Metropolitan University
Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency (JAXA)(ISAS)
Tokyo Metropolitan University
Okayama University of Science
Tokyo Metropolitan University
出版者
宇宙航空研究開発機構(JAXA)
出版者（英）
Japan Aerospace Exploration Agency (JAXA)
雑誌名
宇宙航空研究開発機構研究開発報告: 宇宙科学情報解析論文誌: 第8号
雑誌名(英)
JAXA Research and Development Report: Journal of Space Science Informatics Japan: Volume 8
There are many craters in the moon. Among them are craters having a special structure called 'central peak' (hereinafter referred to as 'the central-peak crater'). This central peak has an important characteristic that substances inside the moon crust are exposed on the moon surface. Therefore, by measuring the surface of the central peak, it is possible to estimate the material of the surrounding inner crust. By analyzing the inner crust, it is expected that estimation of the cause of craters and central peaks, the process of the environment of the moon surface, and crustal deformation of the past. However, except for some famous craters, the investigation has not progressed much. The reason for this is that the discovery of the central peak is based on visual observation of images by experts, so there are few known the central-peak craters. In order to solve this problem, it is necessary to automate the discovery method of the central-peak crater and prepare a catalog that records the position and size of central peaks, thereby greatly increasing the prospecting point candidate of the central-peak crater. Therefore, in this research, the final goal is to create a catalog of the central-peak crater, and for that purpose we propose an automatic discovery method of the central-peak crater. In this research, we use Digital Elevation Model (DEM) of the lunar surface observed by JAXA's lunar orbit satellite 'KAGUYA' to identify the central-peak crater by machine learning and verify its accuracy. Specifically, we first extract craters using a high-speed crater automatic extraction method called 'Rotating Pixel Swapping Method for DEM', label them, and then try to identify the central-peak crater by CNN. As a result, it was impossible to obtain a highly accurate discrimination model that could create the catalog of the central-peak craters, but we could confirm the possibility that CNN is an effective method in the central-peak crater identification.
内容記述
形態: カラー図版あり
内容記述（英）
Physical characteristics: Original contains color illustrations
かぐやDEMを用いた, 機械学習による中央丘クレーター識別
AA1830035001.pdf
(1.64MB)
