High Temporal and Spatial Resolutions of Sea Surface Current from Low-Altitude Remote Sensing

Abstract

Jung, D.; Lee, J.S.; Baek, J.Y.; Nam, J.; Jo, Y.H.; Song, K.M., and Cheong Y.I., 2019. High temporal and spatial resolutions of sea surface current from low-altitude remote sensing. Jung, H.-S.; Lee, S.; Ryu, J.-H., and Cui, T. (eds.), Advances in Remote Sensing and Geoscience Information Systems of Coastal Environments. Journal of Coastal Research, Special Issue No. 90, pp. 282-288. Coconut Creek (Florida), ISSN 0749-0208.The Saemangeum coast in Korea, which is the study area, is divided from the outer side and the inner side of the dyke. Because exchanging seawater are occurred by only two gates (Sinsi and Garyeok), coastal erosion and sedimentation, migration of suspended solids, and chemical component exchange occur extensively along the gates. Therefore, it is critical to observe the movement of seawater flowing through the gates. In this study, an algorithm to extract high-resolution sea surface current information using a Helikite (the compound word for Helium and Kite) and a drone was examined. First, direct georeferencing was performed using Global Positioning System/Inertial Measurement Unit (GPS/IMU) data from the aircraft to assign ground coordinates into the images without a Ground Control Point (GCP). Subsequently, the sea surface current was estimated using a Robust Optical Flow (ROF) algorithm. ROF computes motion by analyzing changes in the brightness values of successive images. The accuracy of the estimated sea surface current was verified by comparing with observations from a surface drifter, and the results showed that the correlation coefficient was 0.57 and the Root Mean Square Error (RMSE) was 8.85 cm/s. Generally, the ROF-based flow rates were underestimated compared to the field measurements. Furthermore, image aliasing occurred due to a 5-s time interval and the currents that had a strong wave frequency could not be completely captured. To fully restore the sea surface current that appeared on the images, it is necessary to obtain the images at various sampling frequencies using video recordings.