Work Description

Title: Machine Learning for Shipwreck Segmentation from Side Scan Sonar Imagery: Dataset and Benchmark Open Access Deposited

AI4Shipwrecks Dataset
h
Attribute Value
Methodology
  • The AI4Shipwrecks dataset contains sidescan sonar images of shipwrecks and corresponding binary labels collected during 2022 and 2023 at the NOAA Thunder Bay National Marine Sanctuary in Alpena, MI. The data collection platform was an Iver3 Autonomous Underwater Vehicle (AUV) equipped with an EdgeTech 2205 dual-frequency ultra-high resolution sidescan sonar and 3D bathymetric system. The labels were compiled from reference labels created by experts in marine archaeology. The intended use of this dataset is to encourage development of semantic segmentation, object detection, or anomaly detection algorithms in the computer vision field. Comparisons of state-of-the-art segmentation networks on our dataset are shown in the paper.
Description
  • The AI4Shipwrecks dataset contains sidescan sonar images of shipwrecks and corresponding binary labels collected during 2022 and 2023 at the NOAA Thunder Bay National Marine Sanctuary in Alpena, MI. The data collection platform was an Iver3 Autonomous Underwater Vehicle (AUV) equipped with an EdgeTech 2205 dual-frequency ultra-high resolution sidescan sonar and 3D bathymetric system. The labels were compiled from reference labels created by experts in marine archaeology. The intended use of this dataset is to encourage development of semantic segmentation, object detection, or anomaly detection algorithms in the computer vision field. Comparisons of state-of-the-art segmentation networks on our dataset are shown in the paper.

  • The file structure is organized as described in the README.txt file, where images in 'images' directories are the waterfall product of sidescan sonar surveys, and images in 'labels' directories are binary representations of expert labels. Images across the 'images' and 'labels' directories are correlated by having identical filenames. In the label images, a pixel value of '0' represents the non-shipwreck/other class and '1' represents the shipwreck class for the correspondingly named image (<wreck_name>_<##>.png) in the images directory.

  • The project webpage can be found at:  https://umfieldrobotics.github.io/ai4shipwrecks/
Creator
Creator ORCID
Depositor
  • anjashep@umich.edu
Contact information
Discipline
Funding agency
  • Other Funding Agency
Other Funding agency
  • National Oceanic and Atmospheric Administration
ORSP grant number
  • AWD018120
Keyword
Resource type
Last modified
  • 01/25/2024
Published
  • 01/25/2024
DOI
  • https://doi.org/10.7302/dmf4-x492
License
To Cite this Work:
Sheppard, A., Sethuraman, A. V., Bagoren, O., Pinnow, C., Anderson, J., Havens, T. C., Skinner, K. A. (2024). Machine Learning for Shipwreck Segmentation from Side Scan Sonar Imagery: Dataset and Benchmark, AI4Shipwrecks Dataset [Data set], University of Michigan - Deep Blue Data. https://doi.org/10.7302/dmf4-x492

Relationships

This work is not a member of any user collections.

Files (Count: 3; Size: 1.14 GB)

Thumbnailthumbnail-column Title Original Upload Last Modified File Size Access Actions
AI4Shipwrecks.zip 2024-01-23 2024-01-23 1.13 GB Open Access
README.txt 2024-01-23 2024-01-23 2.98 KB Open Access
thumbnail.png 2024-01-25 2024-01-25 2.1 MB Open Access

README.txt

Prepared by Anja Sheppard (January 2024)

Title: Machine Learning for Shipwreck Segmentation from Side Scan Sonar Imagery: Dataset and Benchmark

Authors: Advaith V. Sethuraman*, Anja Sheppard*, Onur Bagoren, Christopher Pinnow, Jamey Anderson, Timothy C. Havens, and Katherine A. Skinner

The project webpage can be found at: https://umfieldrobotics.github.io/ai4shipwrecks/

The AI4Shipwrecks dataset contains sidescan sonar images of shipwrecks and corresponding binary labels collected during 2022 and 2023 at the NOAA Thunder Bay National Marine Sanctuary in Alpena, MI. The data collection platform was an Iver3 Autonomous Underwater Vehicle (AUV) equipped with an EdgeTech 2205 dual-frequency ultra-high resolution sidescan sonar and 3D bathymetric system. The labels were compiled from reference labels created by experts in marine archaeology. The intended use of this dataset is to encourage development of semantic segmentation, object detection, or anomaly detection algorithms in the computer vision field. Comparisons of state-of-the-art segmentation networks on our dataset are shown in the paper.

This work is supported by the NOAA Ocean Exploration program under Award #NA21OAR0110196.

The file structure is organized as follows, where images in 'images' directories are the waterfall product of sidescan sonar surveys, and images in 'labels' directories are binary representations of expert labels. In the labels, '0' represents the non-shipwreck/other class and '1' represents the shipwreck class.

root
- test
- images
- _<##>.png
- labels
- _<##>.png
- train
- images
- _<##>.png
- labels
- _<##>.png
- extras
- terrain

The wreck sites have been separated into test and train sets according to the method laid out in further detail in the paper. Further information about the wrecks themselves can be found at: https://thunderbay.noaa.gov/shipwrecks/

The wrecks in train and test are listed below:

train:
- Alpena Steamer*
- Bay City*
- DM Wilson
- DR Hanna
- EB Allen
- Egyptian
- Grecian
- Harvey Bissell*
- Heart Failure
- Isaac M Scott
- Montana
- Oscar T Flint
- Pewabic
- WP Rend

test:
- Artificial Reef
- Barge No 1
- B Franklin**
- Corsair
- Corsican
- James Davidson**
- John F Warner***
- WH Gilbert
- Haltiner Barge
- Lucinda van Valkenburg
- Monohansett
- Monrovia
- Shamrock***
- WP Thew
- Viator

* Alpena Steamer, Bay City, and Harvey Bissell are all found within images tagged "Near Shore"
** B Franklin and James Davidson are both found in images tagged "Davidson"
*** John F Warner and Shamrock are both found in images tagged "Shamrock"

We also provide some terrain-only images, contained within the "terrain" directory under extras. There are 4 sites:
terrain:
- Exploratory A
- Exploratory B
- Exploratory C
- Mischelley Reef

Any questions about the dataset release can be directed to the Field Robotics Group at the University of Michigan, or to Anja Sheppard (anjashep at umich dot edu).

Download All Files (To download individual files, select them in the “Files” panel above)

Total work file size of 1.14 GB may be too large to download directly. Consider using Globus (see below).

Files are ready   Download Data from Globus
Best for data sets > 3 GB. Globus is the platform Deep Blue Data uses to make large data sets available.   More about Globus

Remediation of Harmful Language

The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.