Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Mapping "Brain Terrain" Regions on Mars using Deep Learning

Kyle A. Pearson, Eldar Noe, Daniel Zhao, Alphan Altinok, Alex Morgan

TL;DR

The study addresses the challenge of automating the search for brain terrain on Mars to illuminate recent water-related geomorphology. It introduces a hybrid deep-learning pipeline that uses a Fourier-domain classifier to efficiently prefilter large HiRISE datasets, followed by segmentation to localize brain terrain, enabling rapid processing of the archive. Training data are built via active learning starting from two labeled brain-terrain examples, with a two-class background vs brain-terrain setup and careful handling of class imbalance across ~60,000 images. On ~52,000 images (~28 TB), the approach yields detections in hundreds of images with about 93% accuracy and roughly a 95% reduction in processing time, providing timely guidance for mission operations and landing-site prioritization, with segmentation masks and code released on GitHub for community use.

Abstract

One of the main objectives of the Mars Exploration Program is to search for evidence of past or current life on the planet. To achieve this, Mars exploration has been focusing on regions that may have liquid or frozen water. A set of critical areas may have seen cycles of ice thawing in the relatively recent past in response to periodic changes in the obliquity of Mars. In this work, we use convolutional neural networks to detect surface regions containing "Brain Coral" terrain, a landform on Mars whose similarity in morphology and scale to sorted stone circles on Earth suggests that it may have formed as a consequence of freeze/thaw cycles. We use large images (~100-1000 megapixels) from the Mars Reconnaissance Orbiter to search for these landforms at resolutions close to a few tens of centimeters per pixel (~25--50 cm). Over 52,000 images (~28 TB) were searched (~5% of the Martian surface) where we found detections in over 200 images. To expedite the processing we leverage a classifier network (prior to segmentation) in the Fourier domain that can take advantage of JPEG compression by leveraging blocks of coefficients from a discrete cosine transform in lieu of decoding the entire image at the full spatial resolution. The hybrid pipeline approach maintains ~93% accuracy while cutting down on ~95% of the total processing time compared to running the segmentation network at the full resolution on every image. The timely processing of big data sets helps inform mission operations, geologic surveys to prioritize candidate landing sites, avoid hazardous areas, or map the spatial extent of certain terrain. The segmentation masks and source code are available on Github for the community to explore and build upon.

Mapping "Brain Terrain" Regions on Mars using Deep Learning

TL;DR

The study addresses the challenge of automating the search for brain terrain on Mars to illuminate recent water-related geomorphology. It introduces a hybrid deep-learning pipeline that uses a Fourier-domain classifier to efficiently prefilter large HiRISE datasets, followed by segmentation to localize brain terrain, enabling rapid processing of the archive. Training data are built via active learning starting from two labeled brain-terrain examples, with a two-class background vs brain-terrain setup and careful handling of class imbalance across ~60,000 images. On ~52,000 images (~28 TB), the approach yields detections in hundreds of images with about 93% accuracy and roughly a 95% reduction in processing time, providing timely guidance for mission operations and landing-site prioritization, with segmentation masks and code released on GitHub for community use.

Abstract

One of the main objectives of the Mars Exploration Program is to search for evidence of past or current life on the planet. To achieve this, Mars exploration has been focusing on regions that may have liquid or frozen water. A set of critical areas may have seen cycles of ice thawing in the relatively recent past in response to periodic changes in the obliquity of Mars. In this work, we use convolutional neural networks to detect surface regions containing "Brain Coral" terrain, a landform on Mars whose similarity in morphology and scale to sorted stone circles on Earth suggests that it may have formed as a consequence of freeze/thaw cycles. We use large images (~100-1000 megapixels) from the Mars Reconnaissance Orbiter to search for these landforms at resolutions close to a few tens of centimeters per pixel (~25--50 cm). Over 52,000 images (~28 TB) were searched (~5% of the Martian surface) where we found detections in over 200 images. To expedite the processing we leverage a classifier network (prior to segmentation) in the Fourier domain that can take advantage of JPEG compression by leveraging blocks of coefficients from a discrete cosine transform in lieu of decoding the entire image at the full spatial resolution. The hybrid pipeline approach maintains ~93% accuracy while cutting down on ~95% of the total processing time compared to running the segmentation network at the full resolution on every image. The timely processing of big data sets helps inform mission operations, geologic surveys to prioritize candidate landing sites, avoid hazardous areas, or map the spatial extent of certain terrain. The segmentation masks and source code are available on Github for the community to explore and build upon.
Paper Structure (2 sections, 1 figure)

This paper contains 2 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Observations and Training Data

Figures (1)

  • Figure 1: Decoding times of various JPEG2000 images in the HiRISE archive using a single thread on a 2.4 GHz 8-Core Intel Core i9 CPU. The decoding time improves if the image is opened at a lower resolution. However, a lower resolution may not always retain enough signal to noise for a computer vision algorithm to act reliably. The image size on disk is listed in the plot legend.