Mars Image Content Classification: Three Years of NASA Deployment and Recent Advances

Authors

  • Kiri Wagstaff Jet Propulsion Laboratory, California Institute of Technology
  • Steven Lu Jet Propulsion Laboratory, California Institute of Technology
  • Emily Dunkel Jet Propulsion Laboratory, California Institute of Technology
  • Kevin Grimes Jet Propulsion Laboratory, California Institute of Technology
  • Brandon Zhao Duke University
  • Jesse Cai California Institute of Technology
  • Shoshanna B. Cole Space Science Institute
  • Gary Doran Jet Propulsion Laboratory, California Institute of Technology
  • Raymond Francis Jet Propulsion Laboratory, California Institute of Technology
  • Jake Lee Jet Propulsion Laboratory, California Institute of Technology
  • Lukas Mandrake Jet Propulsion Laboratory, California Institute of Technology

Keywords:

Image Classification, Class Discovery, Classifier Calibration, Mars

Abstract

The NASA Planetary Data System hosts millions of images acquired from the planet Mars. To help users quickly find images of interest, we have developed and deployed content-based classification and search capabilities for Mars orbital and surface images. The deployed systems are publicly accessible using the PDS Image Atlas. We describe the process of training, evaluating, calibrating, and deploying updates to two CNN classifiers for images collected by Mars missions. We also report on three years of deployment including usage statistics, lessons learned, and plans for the future.

Downloads

Published

2021-05-18

How to Cite

Wagstaff, K., Lu, S., Dunkel, E., Grimes, K., Zhao, B., Cai, J., Cole, S. B., Doran, G., Francis, R., Lee, J., & Mandrake, L. (2021). Mars Image Content Classification: Three Years of NASA Deployment and Recent Advances. Proceedings of the AAAI Conference on Artificial Intelligence, 35(17), 15204-15213. Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/17784

Issue

Section

IAAI Technical Track on Highly Innovative Applications of AI