Skip to content

A CNN classifier for NASA HiRISE images of geographical features on Mars.

Notifications You must be signed in to change notification settings

niehusst/HiRISE-Net

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

HiRISE-Net

A personal project to replicate the results of the "Deep Mars" paper by Wagstaff et al. In this project I build a simple forward feed Convolutional Neural Network (CNN) to classify image from the Mars orbital image (HiRISE) labeled data set. Training for 5 epochs with the current model yields about 81% accuracy on the testing data.

Contents:

  • map-proj/: Directory containing individual cropped landmark images
  • labels-map-proj.txt: Class labels (ids) for each landmark image
  • label_data.py: Python dictionary that maps class ids to semantic names
  • deps.txt: Dependencies of this project (that can be pip installled)
  • classifier_model.py: The tensorflow model and data cleaning

Author

Acknowledgements

The HiRISE data used in this project comes from the DOI:

10.5281/zenodo.1048301

Idea for this project and the data originates from the following paper:

Kiri L. Wagstaff, You Lu, Alice Stanboli, Kevin Grimes, Thamme Gowda, and Jordan Padams. "Deep Mars: CNN Classification of Mars Imagery for the PDS Imaging Atlas." Proceedings of the Thirtieth Annual Conference on Innovative Applications of Artificial Intelligence, 2017.

About

A CNN classifier for NASA HiRISE images of geographical features on Mars.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages