Author(s): Bendik Aas; Oskar G. Veggeland; Ekaterina Kim

Linked Author(s):

Keywords: Sea ice; Ice drift; State estimation; Kalman filter; Optical flow; Lukas–Kanade

Abstract: We present a method for tracking and predicting the motion of sea ice floes from ship-based optical cameras. The data processing framework combines optical flow tracking of key points using the Lucas–Kanade method with a Kalman filter for state estimation of individual ice floes. The proposed method was verified using a custom-built simulator and demonstrated promising performance when tested with real-world data. The estimated ice floe trajectories aligned well with the reference RGB imagery from Sentinel-2. Although fine-tuning of the noise parameters for real-world scenarios was beyond the scope of this study, the results indicate that the method captures the general floe motion well. Potential developments toward real-world deployment are also outlined.

DOI:

Year: 2026