It is a powerful tool to count lunar craters for estimating lunar age. It has been up to human eyes to identify and/or count lunar craters with lunar images. However, there are many difficulties in this way due to have potential errors by human eyes. Recently, Christopher(2019) and Silburt et al.(2016) carried to identify the craters at Mars and Moon. Although the results have about 10% error, the cause of error can be clearly defined as using machine. We will try to identify and/or count lunar crates using previous or modified deep-learning model. The data required for learning and testing is WAC/NAC in LROC. We consider to apply to the data of ShadowCam, based on the experience gained through this study, although there are some problems to be solved. The problems; 1) whether or not the model, which is learned with WAC/NAC in LROC, can be applied to the data of ShadowCam, after studying the optical properties of PSR, 2) the distortion of topography on polar region. We expect that there are technological advances for calculating the age of solar system bodies by crater counting using deep-learning. Thanks : This research was conducted by NRF (2018M1A3A3A02065832) support.
展开▼
