Estimate of Coastal Water Depth Based on Aerial Photographs Using a Low-Altitude Remote Sensing System

Abstract

Observing coastal water depths is very important to understand physical processes of the coastal environment and manage coastal resources. In this study, a low-altitude remote sensing system, Helikite, with an optic sensor was used to measure shallow water depths. This system not only enables us to observe the highresolution water depths in the coastal waters in a very short time, but also allow us to estimate water depths of the breaking wave zone where erosion is most active. The “Anmok Beach” study area has undergone severe coastal erosion from typhoons in summer and abnormally high waves in winter. Therefore, aerial photographs were taken and a field water depth survey was carried out at Anmok Beach to estimate water depths. In order to convert aerial orthomosaic images to water depths, the following three steps were undertaken. First, georeferencing was performed via the Oceanic Ground Control Point based on a ground reference point available on the sea surface. Second, error values generated from the aeolian current and sunglint were filtered out from the orthomosaic images. Third, the relationships between in-situ water depth data and brightness values of the orthomosaic image were calculated using the second polynomial regression equation. The obtained regression coefficients (R2) values were 0.90, 0.95, and 0.91 for the R, G, and B-bands, respectively and the corresponding RMSE values were 0.45 m, 0.30 m, and 0.44 m for the R, G, and B-bands, respectively. The accuracy of the G-band based water depth was highest. Using this result, we could estimate the submarine topography of the crescent bar produced by complex coastal currents in the study area. © 2019, Korea Institute of Ocean Science & Technology (KIOST) and the Korean Society of Oceanography (KSO) and Springer Science+Business Media B.V.