UAV video-based estimates of nearshore bathymetry
Athina M.Z. Lange, Julia W. Fiedler, M. A. Merrifield, R. T. Guza
Abstract
Nearshore bathymetry estimated from video acquired by a hovering UAV is compared with ground truth. Bathymetry estimates from the widely used cBathy algorithm are improved by crest tracking (with machine learning-aided annotations) from near breaking through the surf zone. Individual wave crests (distinguished from the breaking wave toe that can move down the wave front face) in video timestacks are determined with a deep-learning neural network and surf zone depth estimates are computed from the wave celerity. Time-2D spatial transforms (cBathy) are used to estimate wave celerity and depth between the surf zone and 10-m depth. Composite profiles (cBathyCT), formed by joining cBathy and crest-tracking solutions near the surf zone seaward edge, based on a newly determined γ(x) parameter, avoid the large cBathy errors associated with the onset of breaking. Including an additional topography survey for the foreshore region, provides full nearshore bathymetry profiles. Incident wave heights were relatively constant on each day but varied over days between 0.55–2.15 m. Averaged over all 17 min hovers and cross-shore transects (112 total), surf zone depths errors were relatively small (average root-mean-square error 〈RMSE〉=0.17m, 〈Bias〉=0.06m) after including a heuristic nonlinear correction to the linear phase speed. Between the seaward surf zone edge and 10 m depth, errors are similar to previous cBathy studies: 〈RMSE〉=0.87m, 〈Bias〉=0.58m with the largest errors in deepest water. Beach profiles were generally similar for all 8 test days, concave up with a slight terrace (no sandbar) and small alongshore depth variations. Accuracy was lower on one transect with a shallow reef.