Towards Reliable Sea Ice Drift Estimation in the Arctic Deep Learning Optical Flow on RADARSAT-2
Daniela Martin, Joseph Gallego
TL;DR
This work tackles the problem of estimating high-resolution sea ice drift from SAR imagery by benchmarking 48 pretrained deep learning optical flow methods on RADARSAT-2 ScanSAR data, with GNSS buoys providing ground-truth trajectories. The authors avoid fine-tuning to assess transferability, revealing that about $EPE$ in the range $6$--$8$ px (≈ $300$--$400$ m) is achievable and that several methods produce dense, consistent drift fields suitable for navigation and climate studies. Qualitative analysis confirms robust regional drift patterns across methods, though some models exhibit magnitude biases and border artifacts, highlighting architecture- and data-dependent trade-offs. The study demonstrates the viability of vision-based optical flow for polar remote sensing and points to domain adaptation, physics-informed constraints, and multi-sensor fusion as promising avenues to further improve performance in real-world Arctic monitoring.
Abstract
Accurate estimation of sea ice drift is critical for Arctic navigation, climate research, and operational forecasting. While optical flow, a computer vision technique for estimating pixel wise motion between consecutive images, has advanced rapidly in computer vision, its applicability to geophysical problems and to satellite SAR imagery remains underexplored. Classical optical flow methods rely on mathematical models and strong assumptions about motion, which limit their accuracy in complex scenarios. Recent deep learning based approaches have substantially improved performance and are now the standard in computer vision, motivating their application to sea ice drift estimation. We present the first large scale benchmark of 48 deep learning optical flow models on RADARSAT 2 ScanSAR sea ice imagery, evaluated with endpoint error (EPE) and Fl all metrics against GNSS tracked buoys. Several models achieve sub kilometer accuracy (EPE 6 to 8 pixels, 300 to 400 m), a small error relative to the spatial scales of sea ice motion and typical navigation requirements in the Arctic. Our results demonstrate that the models are capable of capturing consistent regional drift patterns and that recent deep learning based optical flow methods, which have substantially improved motion estimation accuracy compared to classical methods, can be effectively transferred to polar remote sensing. Optical flow produces spatially continuous drift fields, providing motion estimates for every image pixel rather than at sparse buoy locations, offering new opportunities for navigation and climate modeling.