Published November 9, 2023 | Version v1.0.2
Model Open

Sidescan Sonar Substrate, Depth, and Shadow Models for PINGMapper v2.0.0

Authors/Creators

  • 1. Northern Arizona University

Description

Substrate, depth, and shadow segmentation models implemented in PINGMapper v2.0.0. Details are highlighted in the following manuscript:

Bodine, C. S., Buscombe, D., & Hocking, T. D. (2024). Automated river substrate mapping from sonar imagery with machine learning. Journal of Geophysical Research: Machine Learning and Computation, 1, e2024JH000135. https://doi.org/10.1029/2024JH000135

Models were trained with Segmentation-Gym.

Files

PINGMapperv2.0_SegmentationModelsv1.0.zip

Files (124.9 MB)

Name Size
md5:b33258fd5110bbe5dd3a6ad8e8c2e39e
124.9 MB Preview Download

Additional details

Related works

Is derived from
Software: https://github.com/Doodleverse/segmentation_gym (URL)
Dataset: 10.5281/zenodo.10119320 (DOI)
Is part of
Software: 10.5281/zenodo.10120054 (DOI)
Is published in
Journal article: 10.1029/2024JH000135 (DOI)

Funding

United States Fish and Wildlife Service
Characterizing Gulf Sturgeon Spawning Habitat, Habitat Use and Origins of Juvenile Sturgeon in the Pearl and Pascagoula River Systems F19AC00836