Abstract:Introduction: Habitats with a taphonomic window are high priority targets for astrobiology missions to Mars [1]. However, if life existed on Mars it did not interact with its physical environment in a way that has been obvious to identify from orbit. Studying terrestrial Mars analogs can inform us of the types of features that serve as habitats with a taphonomic window on Earth, and by extension the types of features that should be sought on Mars. But what spatial resolutions are necessary to positively identify such targets? To address the problem of establishing thresholds of identification in spatial resolution, we are assessing the ability of a deep convolutional neural network (CNN) to perform semantic segmentation (classify each pixel) on images of the astrobiologically-relevant Mars-analog environment, Salar de Pajonales (SdP), in Chile (Fig. 1A). Our aim is to discover the resolutions at which the habitats at SdP are no longer identifiable by the CNN to understand if the same types of features could be identified in salt-encrusted basins on Mars. CNNs excel at many tasks related to image analysis, including image classification [2], with modern CNNs capable of outperforming human analysts [2, 3]. Our investigation requires a consistent analysis of the same terrain over many ground sampling distances (GSDs). A single human analyst would become biased after the first classification task, and many human analysts would vary in their ability to classify a scene. Therefore, a CNN is a natural choice to tackle this problem. Here we present initial results of the semantic segmentation of our highest resolution (3 cm/pixel) scene of SdP. In future work, we will evaluate the CNN performance at progressively lower GSDs to establish thresholds of identification for habitats at SdP. Geologic Background and Relevance to Mars: Salar de Pajonales (SdP) is a salt-encrusted basin in the Chilean Altiplano (25°08'29"S, 68°46'20"W, 3547m, Fig. 1A) [4]. Its polyextreme characteristics make it a suitable analog for a post-Noachian martian climate [5]. Our study site within SdP is a gypsum-dominated area that notably hosts ridges and tumuli [6] that serve as endolithic habitats [5], and a “patterned ground” unit that represents a biological soil crust [5] (Table 1, Fig. 1B). Data and Methods: We captured images of our

Planetary Mapping Using Deep Learning: A Method to Evaluate Feature Identification Confidence Applied to Habitats in Mars-Analog Terrain

CONVOLUTIONAL NEURAL NETWORK IMAGE CLASSIFICATION OF MARS-ANALOG TERRAIN: PRELIMINARY RESULTS AND IMPLICATIONS FOR THE SEARCH FOR LIFE ON MARS.

Mapping "Brain Terrain" Regions on Mars using Deep Learning

Deep Learning based Semantic Segmentation for Mars Rover Terrain Classification

Research and ground-based validation on Mars rover localization based on multi-level images matching

Automatic Data Processing for Space Robotics Machine Learning

Deep Mars: CNN Classification of Mars Imagery for the PDS Imaging Atlas

Multi-mission Terrain Classifier for Safe Rover Navigation and Automated Science

Generative 3D Reconstruction of Martian Surfaces Using Monocular Images

MarsMapNet: A Novel Superpixel-Guided Multiview Feature Fusion Network for Efficient Martian Landform Mapping

Highly Accurate Visual Method of Mars Terrain Classification for Rovers Based on Novel Image Features

Self-Supervised Learning to Guide Scientifically Relevant Categorization of Martian Terrain Images

An automatic classification method for mapping Martian landforms

Non-myopic Planetary Exploration Combining In Situ and Remote Measurements

Spatial Spectroscopic Models for Remote Exploration

Terrain Classification Using Mars Raw Images Based on Deep Learning Algorithms with Application to Wheeled Planetary Rovers

Signatures of Life Detected in Images of Rocks Using Neural Network Analysis Demonstrate New Potential for Searching for Biosignatures on the Surface of Mars

Dense Feature Matching for Hazard Detection and Avoidance Using Machine Learning in Complex Unstructured Scenarios

AI4MARS: A Dataset for Terrain-Aware Autonomous Driving on Mars

Mobile-DeepRFB: A Lightweight Terrain Classifier for Automatic Mars Rover Navigation

LARGE SCALE MARS MAPPING AND ROVER LOCALIZATION USING DESCENT AND ROVER IMAGERY