Automated precision counting of small lunar craters - A broader view

Abstract

An automated system for counting small craters at lunar landing sites has been expanded to include selected locations on the Moon's far side and high latitudes. Highland areas were generally excluded from this new survey, because sloping terrain can have a significant effect on crater lifetimes. But some locations on highland light plains units were analysed, as were some young Copernican age crater ejecta blankets and impact melts.This study focuses on craters having diameters between 2.5 m and 50 m, for which counts are so high that they are best plotted as Crater Size Frequency Histograms (CSFHs), rather than the conventional cumulative plots (CSFDs). This histogram representation enables the finest details of the plots to be compared across sites. Using multiple precisely aligned images of each site enables false negatives to be included and false positives rejected, ensuring that no visually recognisable craters are missed, a capability that is not possible when analysing single images.There are some very distinct differences between the curves at different sites. Based on the known ages of boulder tracks and counts of craters formed over the last 50 years, the largest craters within the size range up to 50 m are thought to be no more that a few hundred million years old. A model for the evolution of small crater populations on an absolute time scale is presented and the absolute ages of young features are estimated.There is no evidence in the data for secondary cratering, nor is there any clear variation of counts and slopes with longitude or latitude, so other possible explanations for the variabilities between sites must be found.The results obtained by this project should prove useful for modelling crater formation and destruction and also as training data for machine learning systems.