Surveying a Floating Iceberg With the USV SEADRAGON

• Published in: Frontiers in Marine Science
• Main Authors: Zhou, Mingxi, Bachmayer, Ralph, DeYoung, Brad
• Format: Text
• Language: unknown
• Published: DigitalCommons@URI 2021
• Subjects: Ice Sheet; Ice Shelves
• Online Access: https://digitalcommons.uri.edu/gsofacpubs/815; https://doi.org/10.3389/fmars.2021.549566; https://digitalcommons.uri.edu/context/gsofacpubs/article/1783/viewcontent/Zhou_M_SurveyingFloating_2021.pdf

The calving, drifting, and melting of icebergs has local, regional, and global implications. Besides the impacts to local ecosystems due to changes in seawater salinity and temperature, the freshwater influx and transport can have significant regional effects related to the ocean circulation. The increased influx of freshwater ice due to increase calving from ice shelves and the destabilization of the continental ice sheet will affect sea levels globally. In addition, drifting icebergs pose threats to offshore operations because they could damage offshore installations, e.g., pipelines and subsea manifolds, and interrupt marine transportation. Iceberg drift and deterioration models have been developed to better predict climate change and protect offshore operations. Iceberg shape is one of the most critical parameters in these models, but it is challenging to obtain because of iceberg movement caused by winds, waves, and currents. In this paper, we present an algorithm for iceberg motion estimation and shape reconstruction based on in-situ point cloud measurements. The algorithm is developed based on point cloud matching strategies, policy-based optimization, and Kalman filtering. A down-sampling method is also integrated to reduce the processing time for possible real-time applications. The motion estimation algorithm is applied to a simulated data set and field measurements collected by an Unmanned Surface Vehicle (USV) on a free-floating, translating, and rotating, iceberg. In the field data, the above-water iceberg surface was measured with a scanning LIDAR, while the below-water portion (0–50 m) was profiled using a side-looking multi-beam sonar. When applying the motion estimation algorithm to these two independent point cloud measurements collected by the two sensing modalities, consistent iceberg motion estimates are obtained. The resulting motion estimates are then used to reconstruct the iceberg shape. During the field experiment, additional oceanographic measurements, such as temperature, ocean ...