Surveying Drifting Icebergs and Ice Islands: Deterioration Detection and Mass Estimation with Aerial Photogrammetry and Laser Scanning

Abstract

Icebergs and ice islands (large, tabular icebergs) are challenging targets to survey due to their size, mobility, remote locations, and potentially difficult environmental conditions. Here, we assess the precision and utility of aerial photography surveying with structure-from-motion multi-view stereo photogrammetry processing (SfM) and vessel-based terrestrial laser scanning (TLS) for iceberg deterioration detection and mass estimation. For both techniques, we determine the minimum amount of change required to reliably resolve iceberg deterioration, the deterioration detection threshold (DDT), using triplicate surveys of two iceberg survey targets. We also calculate their relative uncertainties for iceberg mass estimation. The quality of deployed Global Positioning System (GPS) units that were used for drift correction and scale assignment was a major determinant of point cloud precision. When dual-frequency GPS receivers were deployed, DDT values of 2.5 and 0.40 m were calculated for the TLS and SfM point clouds, respectively. In contrast, values of 6.6 and 3.4 m were calculated when tracking beacons with lower-quality GPS were used. The SfM dataset was also more precise when used for iceberg mass estimation, and we recommend further development of this technique for iceberg-related end-uses.

Mapping the underside of an iceberg with a modified underwater glider

“…In contrast, the point cloud collected on June 16 was found has a mean southward translation about 26 m. Next, we applied multiscale model‐to‐model could comparison (M3C2; Lague, Brodu, & Leroux, ) to compute the linear distance between corresponding points in the different point clouds. The method was used for above‐water iceberg point cloud comparison and deterioration estimates (Crawford, Mueller, & Joyal, ). The resulting M3C2 distance between individual point clouds shown in Figure is presented in Figure where 95% of the distances found were less than 10 m for individual comparisons.…”

Sea-ice topographic surveying using Structure-from-Motion photogrammetry conducted from small UAVs

“…Previous research has successfully used SfM surveys in cold regions despite these challenges. This has been demonstrated, for example, by manned aircraft-borne SfM surveys of low-contrast surfaces (snow) by Nolan et al (2015) or of non-stationary targets (ice islands) by Crawford et al (2018) and by successful sUAV-SfM surveys of glaciers by Ryan et al (2015). Some uses of SfM for surveys of sea-ice have been reported from both manned aircraft (Steer, 2017) and sUAVs (Eltoft et al, 2015;Lean, 2012;Saiet et al, 2015;Saiet, 2015).…”

“…This increased flexibility and accessibility is key to understanding the appeal of SfM. It is increasingly used for general geoscience surveys on firm land (Brunier et al, 2016;Dietrich, 2016;Ely et al, 2016;Lucieer et al, 2014;Mancini et al, 2013;Smith and Damià, 2015;Westoby et al, 2012) and it is also used for snow-depth surveys (Nolan et al, 2015) and ice island surveys (Crawford et al, 2018). However, it has been rarely used to survey sea-ice, although a few examples can be found (ACUASI, 2015;Dammann et al, 2018;Saiet et al, 2015;Saiet, 2015;Steer, 2017).…”

“…GPS receivers are commonly used to scale and/or georeference SfM surveys (Crawford et al, 2018;Nolan et al, 2015;Smith et al, 2016). The mean accuracy of consumer-grade, single-frequency (L1), GPS receivers, which uses code (binary signals) pseudorange measurements for triangulation, is about 5 metres (van Diggelen and Enge, 2015).…”

“…SfM can accommodate lens distortions, common in lower-end cameras, as these distortions will be minimized in further steps, enabling low capturing costs (Mosbrucker et al, 2016); and 3. Environmental parameters (Crawford et al, 2018;Nolan et al, 2015), such as atmosphere transparency, cloud cover, sun angle, sun intensity, ground reflectivity, ground homogeneity and GCPs.…”