USING OPTICAL FLOW TO ESTIMATE GLACIER DISPLACEMENTS IN THE SOUTH PATAGONIA ICEFIELD

摘要

In this work, the feasibility of using optical flow as a possible solution to obtain accurate movement data at pixel level to derive ice velocities in a glacier was investigated. The study of glacier dynamics requires the accurate mapping of surface velocities that vary along the glacier, following complex patterns defined by stress and strain rate distributions. In order to obtain a dense and accurate grid of ice velocities by optical flow algorithm as the result of the apparent movement pattern between objects, we carried out a test of the proposed large displacement optical flow (LDOF) method in Viedma Glacier, located at the Parque Nacional Los Glaciares, South Patagonia Icefield, Argentina. We collected a monoscopic image terrestrial sequence (time lapse), acquired by a calibrated camera; images were taken every 24 hour from April 2015 until April 2016, a total of 346 days. The Correlated Image Filter (CIF) process was applied to avoid and minimize errors due to the significant changes in lighting, shadows, clouds and snow that allowed to select a sequence of correlated image pairs of comparable radiometric characteristics. The results show a strong flow field in the direction of the glacier movement with acceleration in the terminus. In addition, the errors between different images pairs were analyzed, and the matches generally appear to be adequate, although some areas show random gross errors related to the presence of substantial differences in lighting. These errors were minimized by averaging the image sequence based on seasons, which yielded better results. In summary, the LDOF method applied to terrestrial time lapse data can provide a fairly good solution to detect large daily changes in the glacier; typically, at sub-pixel level.