Glacier4D: Glacier margin change monitoring on the basis of multi-temporal 3D point clouds

Project title

Glacier4D: Glacier margin change monitoring on the basis of multi-temporal 3D point clouds

Funding / Duration

German Research Foundation (DFG) / 3 years (Start: 01/2021)
Project number: 436500674 (-> GEPRIS)

Description

In response to global climate change, glaciers worldwide indicate significant and partly massive changes, primarily characterised by glacier front retreat and thinning. Both of these correspond to a loss of ice mass. Glacier monitoring typically focuses on measuring the retreat and advance as an indicator of climatic change, which can be obtained from monoscopic satellite, aerial or terrestrial imagery. While the frontal position of glaciers calving into seas and lakes can usually be well identified in images, the frontal position of debris-covered or rock glaciers is often difficult to detect and is one of the main error sources in glacier mapping.

The project aims to develop photogrammetric methods for the reliable mapping of debris-covered glacier fronts. The main focus is on high mountain glaciers, as they interact very sensitively with climate change and show excessive retreat in some areas. Known methods employ image segmentation and classification techniques using different spectral bands, which are not very effective for debris-covered glaciers.

A solution can be found in the spatio-temporal extension of the 2D image space and thus in a 4D approach: While it is often almost impossible to delineate the margin of a debris-covered glacier in 2D images, height changes induced by glacier retreat (or advance) can be reliably detected in multi-temporal 3D surface models generated from stereoscopic image pairs, and thus the relative margin position. In order to identify stagnant glacier outlines, the approach is extended by tracking unique features at the surface to detect margins by segmenting static and dynamic parts of the glacier.

The methods will be developed with respect to different scales ranging from mid-resolution satellite imagery to high-resolution satellite imagery, aerial imagery and UAV (unmanned aerial vehicle) imagery. Three study areas with different spatial extents and characteristics are currently planned for validation, although the applicability of the methods is not limited to these:

  • Boverbrean, Jotunheimen National Park (Norway)
  • Bellinghausen Dome, King George Island (Antarctica)
  • Aletsch Glacier, Alps (Switzerland)
  • Skeiðarárjökull, Vatnajökull (Iceland)

Virtual flyover using a 3D surface model of Bellinghausen Dome, King George Island (Antarctica) from UAV images (03/2019) © IPF

Privacy Policy
Debris-covered margin of Bellingshausen Dome (oblique UAV image (DJI Mavic Air), 03/2019)
Debris-covered margin of Bellingshausen Dome (oblique UAV image (DJI Mavic Air), 03/2019)

Debris-covered margin of Bellingshausen Dome (oblique UAV image (DJI Mavic Air), 03/2019) © IPF

Change detection at Bellinghausen Dome, King George Island, using two 3D surface models from UAV images (Temporal offset: 03/2019 | 02/2020). Mapping of significant decreases of debris-covered glacier and snow areas in metres
Change detection at Bellinghausen Dome, King George Island, using two 3D surface models from UAV images (Temporal offset: 03/2019 | 02/2020). Mapping of significant decreases of debris-covered glacier and snow areas in metres

Change detection at Bellinghausen Dome, King George Island, using two 3D surface models from UAV images (Temporal offset: 03/2019 | 02/2020). Mapping of significant decreases of debris-covered glacier and snow areas in metres. © IPF

Change detection at Bellinghausen Dome, King George Island, using two 3D surface models from UAV images (Temporal offset: 03/2019 | 02/2020). Co-registration of the 3D point clouds and change detection in a profile section (unit: metres).
Change detection at Bellinghausen Dome, King George Island, using two 3D surface models from UAV images (Temporal offset: 03/2019 | 02/2020). Co-registration of the 3D point clouds and change detection in a profile section (unit: metres).

Change detection at Bellinghausen Dome, King George Island, using two 3D surface models from UAV images (Temporal offset: 03/2019 | 02/2020). Co-registration of the 3D point clouds and change detection in a profile section (unit: metres). © IPF

Related papers

Subject to change without notice!

7 Entries

2024

Logo Research Information System This information is provided by the FIS.

Contact

About this page

Melanie Elias
Last modified: Jun 27, 2024