An Adaptive Stacking-OF Method for Extracting Glacier Velocity Field Using Optical Remote Sensing Datasets

The optical flow (OF) algorithm has been employed in the extraction of glacier flow velocity fields due to its efficient data processing capabilities. However, rapid changes on the glacier surface result in decreased similarity in remote sensing image series, leading to an increase in outliers within OF-driven offset stacks. This limits the accuracy of flow velocity field extraction. In this letter, we propose a novel adaptive Stacking-OF method designed to accurately compute the glacier flow velocity field. Our approach introduces a crucial temporal baseline criterion to define an effective subset of OF-driven offsets that are used as input for subsequent operations. We systematically filter outliers from offset stacks obtained by iterative regression analysis. Finally, a Stacking operation is employed to calculate the glacier flow velocity field based on the refined multitemporal offset series. We conducted tests on Gongba Glacier, using Landsat 8 and Sentinel-2 satellite images from 2018 to 2020. The experimental results demonstrate the excellent advantages of our proposed method in refining the stacked offset data. Consequently, the retrieved glacier flow velocity field is rendered more precise and reliable.