A Labeled Image Dataset for Deep Learning-Driven Rockfall Detection on the Moon and Mars

by Bickel, Valentin & 2 more Authors | More information
Information

Background: The term rockfall describes the rapid displacement of a large, usually meter-sized block of rock down-slope, triggered by, for example, endogenic or exogenic events like impacts, quakes or rainfall. In a remote sensing context, the term rockfall is also being used to describe the characteristic geomorphic deposit of a rockfall event that can be identified from an air- or space-borne perspective, i.e., the combination of a displaced boulder and the track it carved into the slope substrate while bouncing, rolling, and sliding over the surface (also called `boulder with track' or `rolling boulder'). In planetary science, the spatial distribution and frequency of rockfalls provide insights into the global erosional state and activity of a planetary body while their tracks act as tools that allow for the remote estimation of the surface strength properties of yet unexplored regions in preparation of future ground exploration missions, such as the lunar pyroclastic, polar sunlit and permanently shadowed regions of the Moon. Due to their small physical size (meters), the identification and mapping of rockfalls in planetary satellite imagery is challenging and very time-consuming, however. For this reason, Bickel et al. (2018) and Bickel et al. (2020) trained convolutional neural networks to automate rockfall mapping in lunar and martian satellite imagery. Parts of the unpublished datasets used for earlier work have now been complemented with newly labeled data to create a well-balanced dataset of 2,822 lunar and martian rockfall labels (which we call `RMaM-2020' --- [R]ockfall [Ma]rs [M]oon [2020], 416 MB in total, available here) that can be used for deep learning and other data science applications. Here, balanced means that the labels have been derived from imagery with a wide and continuous range of properties like spatial resolution, solar illumination, and others. So far, this dataset has been used to analyze the benefits of multi-domain learning on rockfall detector performance (Mars & Moon vs. Moon-only or Mars-only), but there are numerous other (non-planetary science) applications such as for featurization, feature or target recognition (aircraft/spacecraft autonomy), and data augmentation experiments.
				
Authors Valentin Bickel (1) , Lukas Mandrake (2), Gary Doran (2) Affiliations 1. Planets and Comets, MPI for Solar System Research, Justus von Liebig Weg 3
2. Machine Learning and Instrument Autonomy, NASA Jet Propulsion Laboratory, 4800 Oak Grove Drive
Cite as Bickel, Valentin, Mandrake, Lukas, Doran, Gary. (2021). A Labeled Image Dataset for Deep Learning-Driven Rockfall Detection on the Moon and Mars. Max Planck Society. https://edmond.mpdl.mpg.de/imeji/collection/DowTY91csU3jv9S2 Study Type(s) other
Keywords
rockfall, Mars, Moon, dataset, object detection
Project
RMaM-2020
Journal
Frontiers in Remote Sensing
Article DOI 10.3389/frsen.2021.640034 Geolocation – Place
Mars, Moon
Number of items in this collection 1 Creation date Wed Feb 10 10:04:16 CET 2021 Last modification date Wed Feb 10 10:50:12 CET 2021 Date of publication Wed Feb 10 10:50:12 CET 2021 Permalink https://edmond.mpdl.mpg.de/imeji/collection/DowTY91csU3jv9S2 QR-code
Items
1 Item
Sort by
Ascending Descending
View options
Thumbnails List
  • RMaM-2020.zip
    File name
    RMaM-2020.zip
    Collection
    A Labeled Image Dataset for Deep Learning-Driven Rockfall Detection on the Moon and Mars
1
of 1
 
Collection is published. Please choose a license.

Cancel
Drag files here...
Calculating...
Calculating...
Starting...
Are you sure you want to delete all items?
Are you sure you want to publish this collection?
Are you sure you want to delete this collection?

All items and subcollections contained in this collection will be deleted.

Are you sure you want to discard this collection?

All items and subcollections contained in this collection will be deleted. Only the metadata will be still available.

By subscribing to this collection you will receive automatic e-mail notification if items are changed or new items are created. Are you sure you want to subscribe to this collection?
Do you really want to unsubscribe from this collection?
By continuing you authorize the Max Planck Digital Library to create a DOI for this collection. The collection URL and metadata will be transferred to the TIB Hannover and made publicly accessible.
The DOI registration is irreversible, therefore please use this function with care and only if you are aware of the consequences, see http://doi.mpdl.mpg.de/faq for further details.
By continuing you authorize the Max Planck Digital Library to reserve a DOI for this collection. The reserved DOI is in draft state, not yet activated and thus not yet resolving to the collection URL.
Only when the collection gets published, the DOI will resolve to the collection URL and be findable for search engines.
The DOI registration is irreversible, therefore please use this function with care and only if you are aware of the consequences, see http://doi.mpdl.mpg.de/faq for further details (here: commitment to make landing page and content of the collection publicly available in due course).
The draft DOI can already be communicated to e.g. journals.

Move to a collection Cancel

Are you sure you want to discard the selected items? All files will be deleted and only the metadata will be still available.

Are you sure you want to delete the selected items?
Move to a collection Cancel