Research Update | New Progress in High-Resolution Underwater Terrain Reconstruction near China’s Qinling Station in Antarctica

Research Background

Against the backdrop of environmental change in Antarctica, high-resolution nearshore bathymetric terrain provides fundamental data for understanding ice–ocean interactions and also offers important support for environmental surveys and the safe operation and maintenance of facilities around Antarctic research stations. However, conventional underwater surveying methods are costly and inefficient in polar shallow-water environments. As an efficient and low-cost technique, low-altitude Uncrewed Aerial Vehicle (UAV)-based Structure-from-Motion (SfM) photogrammetry provides new opportunities for coastal geomorphological research.
To address the challenges posed by water refraction, floating ice, and large boulders to underwater terrain extraction using UAV photogrammetry in Antarctic nearshore shallow waters, the teams led by Professors Wei Feng and Lei Zheng from the School of Geospatial Engineering and Science, Sun Yat-sen University, selected Seal Bay near China’s Qinling Station as a case study to explore the application of UAV photogrammetry for rapid shallow-water terrain mapping, providing a new technical solution for Antarctic nearshore environmental monitoring and the operation and maintenance of research stations.

Paper Overview

Taking Seal Bay near China’s Qinling Station as the study area (Fig. 1), this study proposed a workflow for nearshore bathymetric terrain extraction based on UAV-based SfM photogrammetry (Fig. 2) and evaluated its applicability.
High-overlap UAV aerial images of Seal Bay acquired during the 38th Chinese Antarctic Research Expedition were used to reconstruct a high-density underwater point cloud through SfM-MVS (Multi-View Stereo). To address the refraction effect at the air–water interface, small-angle refraction correction and multi-view refraction correction were applied. Cloth Simulation Filtering (CSF) was then employed for ground filtering of the point cloud, successfully separating underwater terrain points from non-terrain points such as floating ice and large boulders. An underwater Digital Terrain Model (DTM) was subsequently generated (Fig. 3).
The results were validated using in situ water-depth measurements collected by an Uncrewed Surface Vehicle (USV). Under nadir-view photography conditions, the small-angle refraction correction achieved slightly higher bathymetric accuracy than the multi-view refraction correction, with an R² of 0.83 and an RMSE of 0.43 m. The study also found that bathymetric errors for both refraction correction methods increased with water depth.
CSF effectively separated underwater terrain information from non-terrain features such as floating ice and large boulders, achieving classification accuracies of more than 97% under different terrain-slope conditions.
The study demonstrates that UAV-based SfM photogrammetry combined with refraction correction and ground filtering can effectively extract bathymetric terrain in Antarctic nearshore shallow waters. A high-quality initial point cloud is fundamental to subsequent processing; refraction correction can effectively reduce bathymetric errors caused by water refraction; and CSF can effectively remove floating ice and large boulders.
In the future, bathymetric terrain extraction accuracy and the applicability of the method could be further improved by optimizing UAV image acquisition strategies, incorporating more oblique imagery, and improving refraction correction methods. These advances could provide more reliable terrain data for Antarctic nearshore environmental monitoring, the operation and maintenance of research stations, and studies of ice–ocean interactions.

Fig. 1. (A) Location of the study area; (B) field photograph of the Xuelong unloading supplies at Qinling Station via a barge; (C) overview of Seal Bay (Photo by Lei Zheng, 2022).

Fig. 2. Technical workflow for underwater terrain extraction at Qinling Station, Antarctica. DSM represents digital surface model; DTM represents digital terrain model.

Fig. 3. Nearshore underwater terrain at Qinling Station, Antarctica. (A) Underwater DSM before refraction correction; (B) underwater DSM after refraction correction; (C) underwater DTM after ground filtering. Black circles, red rectangles, and dashed rectangles indicate large boulders, floating ice, and filtering anomalies, respectively.

Publication Information

The study was published online in Journal of Remote Sensing in September 2026 under the title “UAV-Based SfM Photogrammetry for Bathymetric Terrain Extraction in Antarctic Shallow Bays near China's Qinling Station.”
Jinchen He, a doctoral student with the Gravity Remote Sensing and Navigation Team, is the first author. Professor Wei Feng from the team and Professor Lei Zheng from the Polar and Ocean Remote Sensing Team are the corresponding authors. Other co-authors include Professor Xiao Cheng, Associate Professor Shuhang Zhang, doctoral student Xiaodong Cui, and postdoctoral researcher Bingtao Chang from Sun Yat-sen University, as well as Senior Engineer Jie Li from the Polar Research Institute of China.
The research was supported by the Discipline Breakthrough Precursor Project of the Ministry of Education of China (JYB2025XDXM803), the Open Fund of Key Laboratory of Marine Environmental Survey Technology and Application, Ministry of Natural Resources (MESTA-2022-B004), and the National Natural Science Foundation of China (42422606).
Citation:
He, J., S. Zhang, W. Feng*, L. Zheng*, X. Cui, B. Chang, J. Li, and X. Cheng. UAV-Based SfM Photogrammetry for Bathymetric Terrain Extraction in Antarctic Shallow Bays near China's Qinling Station. Journal of Remote Sensing, 2026. doi:10.34133/remotesensing.1071.