UAV-Based SfM Photogrammetry for Bathymetric Terrain Extraction in Antarctic Shallow Bays Near China’s Qinling Station

High-resolution nearshore bathymetric terrain is the baseline data for understanding the ice–ocean interaction in the Antarctic and is crucial for the safe operation and maintenance of Antarctic research stations. Particularly, for shallow-water topography, traditional surveying methods are inefficient and costly, while satellite remote sensing techniques are limited by spatial resolution. Uncrewed aerial vehicle (UAV)-based structure-from-motion (SfM) photogrammetry, as a high-efficiency and low-cost technique, provides new opportunities for coastal geomorphic research. However, refraction effect, floating ices, and large gravels pose challenges to bathymetric terrain extraction from UAV-based SfM photogrammetry in Antarctic gravel-bed bays. This study takes Seal Bay near the Qinling Station of China as an example and presents the workflow and applicability assessment for extracting bathymetric terrain of gravel-bed bays using UAV-SfM photogrammetry. This workflow utilizes UAV-captured overlapping images to reconstruct the initial dense point cloud, and accurate bathymetric digital terrain models are obtained by combining refraction correction and ground filtering methods. Validation against in situ uncrewed surface vehicle measurements shows that the R 2 (coefficient of determination) of small-angle refraction correction method is 0.83, with an RMSE (root mean square error) of 0.43 m, under nadir viewing angle conditions. Meanwhile, the classification accuracy of underwater terrain extraction with cloth simulation filtering for removing floating ices and gravels exceeds 97%. This study provides a cost-effective strategy for bathymetric terrain extraction that can easily be applied in other shallow gravel-bed bays along the Antarctic coastlines.