Coastal bathymetry is a fundamental component in numerous research and engineering applications, however obtaining such data in detail for large areas is rather complicated due to several natural and technical limitations. In order to overcome these difficulties, we developed a pipeline of novel methods based on drone-imagery for producing high resolution, accurate coastal bathymetric data. As a result, this study was based on combination of state-of-the-art, high resolution datasets, including drone-based imagery and reference bathymetric points using a sonar sensor mounted on a remote-controlled unmanned surface vehicle (USV). Our approach relies on fusion of two conceptually different yet complementary methods for bathymetry estimation, a geometric one and a spectral one using deep learning. In the first approach, we developed a novel structure-from-motion technique, that natively incorporates the correct geometrical optics accounting for water refraction. In the second method we built upon the reconstructed surface of the geometric method together with the reference bathymetric data for training a deep convolutional network (CNN) using a set of spectral features from the RGB imagery. The CNN produced high resolution coastal bathymetry with vertical accuracy varying in the decimeter scale. The main advantage of this approach is that it exploits both the spectral and the multi-view aspects of drone imagery which function complementary to each other. The geometric method yields accurate 3D bathymetry over any kind of seafloor that shows sufficient texture on the images and where seafloor texture is absent, spectral information is utilized for harmonizing the bathymetry surface for the entire scene. Finally, this study demonstrates that modern, unmanned platforms can perform accurate coastal bathymetry mapping far more efficiently than traditional boat surveys, although ideal sea-state conditions are required for obtaining imagery data with optimal quality. This work is part of the ACTYS project (https://actys.ims.forth.gr/) that has received funding from FORTH-Synergy grant.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.