Deriving Surface Ages on Mars Using Automated Crater Counting

Abstract

Impact craters on solar system bodies are used to determine the relative ages of surfaces. The smaller the limiting primary crater size, the higher the spatial resolution in surface/resurfacing age dating. A manually counted database (Robbins & Hynek, 2012, https://doi.org/10.1029/2011JE003966) of >384,000 craters on Mars >1 km in diameter exists. But because crater size scales as a power law, the number of impact craters in the size range 10 m to 1 km is in the tens of millions, a number making precise analysis of local variations of age, over an entire surface, impossible to perform by manual counting. To decode this crater size population at a planetary scale, we developed an automated Crater Detection Algorithm based on the You Only Look Once v3 object detection system. The algorithm was trained by annotating images of the controlled Thermal Emission Imaging System daytime infrared data set. This training data set contains 7,048 craters that the algorithm used as a learning benchmark. The results were validated against the manually counted database as the ground truth data set. We applied our algorithm to the Thermal Emission Imaging System global mosaic between ±65° of latitude, returning a true positive detection rate of 91% and a diameter estimation error (~15%) consistent with typical manual count variation. Importantly, although a number of automated crater counting algorithms have been published, for the first time we demonstrate that automatic counting can be routinely used to derive robust surface ages.

The Boulder Population of Asteroid 4 Vesta: Size‐Frequency Distribution and Survival Time

“…Despite our best efforts, systematic errors may result from mapping being performed by a single individual. Such bias may be avoided by the use of automated boulder detection algorithms, such as one recently developed for crater counting purposes (Benedix et al, 2020). The location of boulders (and craters) is defined in the Claudia coordinate system (Russell et al, 2012).…”

Model Age Derivation of Large Martian Impact Craters, Using Automatic Crater Counting Methods

“…The Crater Detection Algorithm (CDA) used to identify craters is described in detail by Benedix et al. (2020). This Convolutional Neural Network (CNN) was originally trained on the THEMIS Day IR mosaic where 1,762 craters from the Robbins and Hynek (2012) have been identified and manually cleaned to select the most identifiable impact craters.…”

“…This Convolutional Neural Network (CNN) was originally trained on the THEMIS Day IR mosaic where 1,762 craters from the Robbins and Hynek (2012) have been identified and manually cleaned to select the most identifiable impact craters. Applied to the THEMIS Day IR mosaic between 45 degrees of North and South, our algorithm produces a true positive detection rate of 86% for craters larger than 1 km in diameter (Benedix et al., 2020). However, the identification of smaller craters require higher resolution data set such as HiRISE (High‐Resolution Imagery System Experiment) or CTX (Context Camera).…”

“…While irregularly shaped craters (highly elliptical, rim scalloped shape) are mostly missed by the previous version of the CDA (Benedix et al., 2020), circular secondary craters produced by high angle ejecta (Zhou et al., 2015) are detected. In this paper, we present the results of our CDA, retrained on high‐resolution imagery datasets.…”

“…However, manual counting is a tedious task, only practical for small areas and/or very young surfaces, as it is greatly affected by human ability and submitted to biases in the recognition of impact structures and the measurement of their diameter (Robbins et al., 2014). In a previous paper (Benedix et al., 2020), we demonstrated our ability to automatically count kilometer‐sized impact craters and derive consistent surface ages using a Crater Detection Algorithm (CDA), specifically trained to detect impact structures on the surface of Mars using optical imagery. However, the dating of young or small geological units requires counting craters smaller than 1 km in diameter, which are naturally more numerous (e.g., Fassett, 2016; W. K. Hartmann, 2005; W. K. Hartmann & Neukum, 2001).…”