Automatic detection of secondary craters and mapping of planetary surface age based on lunar orbital images

Abstract

Ages of planetary surfaces are typically obtained by manually determining the impact crater size-frequency distribution (CSFD) in spacecraft imagery, which is a very intricate and time-consuming procedure. In this work, an image-based crater detection algorithm that relies on a generative template matching technique is applied to establish the CSFD of the floor of the lunar farside crater Tsiolkovsky. The automatic detection threshold value is calibrated based on a 100 km2 test area for which the CSFD has been determined by manual crater counting in a previous study. This allows for the construction of an age map of the complete crater floor. It is well known that the CSFD may be affected by secondary craters. Hence, our detection results are refined by applying a secondary candidate detection (SCD) algorithm relying on Voronoi tessellation of the spatial crater distribution, which searches for clusters of craters. The detected clusters are assumed to result from the presence of secondary craters, which are then removed from the CSFD. We found it favourable to apply the SCD algorithm separately to each diameter bin of the CSFD histogram. In comparison with the original age map, the refined age map obtained after removal of secondary candidates has a more homogeneous appearance and does not exhibit regions of spuriously high age resulting from contamination by secondary craters.