High Resolution 3D Geological Mapping Using Structure-from-Motion Photogrammetry in the Deep Ocean Bathyal Zone
Abstract
Marine geological studies use detailed maps for quantitative assessment of sea-floor structures that cannot otherwise be visualized. Common methods for imaging the ocean floor at bathyal depths (>1000 m) rely on shipboard acoustic methods, primarily multibeam bathymetry and side-scan sonar/acoustic backscatter that have resolutions from several hundred meters to a few meters. Video photography by a Remotely Operated Vehicle (ROV), in contrast, provides images with centimeter-scale spatial resolution as well as color and texture information. However, creating quantitative maps from an ROV video stream remains a challenge, because of the need to adjust for constantly changing illumination with a limited depth of field and the lack of precise location of the ROV. We modified Agisoft Metashape Pro© software typically used for terrestrial photogrammetry from a moving camera (e.g., on a drone) to the subaqueous setting to create 3D maps of submarine structures within the bathyal zone. We specifically applied the structure-from-motion (SfM) technique to a 2013 dive at depths of 1100-2200 m along the edge of the Mona Passage by the ROV Hercules of the Ocean Exploration Trust. Multibeam bathymetry and seismic reflection data documented a ~10 km3 landslide deposit there, which located near seafloor communication cables cut during the 1918 7.5 ML earthquake. These observations led Lopez, ten Brink, and Geist (Mar. Geol., 254, 35-46, 2008) to suggest a landslide source for the fatal tsunami that followed the earthquake. Examination of outcrops along the bottom and side walls of the slides coupled with grab samples and shallow cores will allow us to determine whether the landslide played a role in the tsunami generation. This research provides the ability to map the seafloor with high-resolution imagery and demonstrates the capability of extracting useful bathymetric data from ROV video.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2021
- Bibcode:
- 2021AGUFM.T35B0205D