EGU General Assembly 2020
© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

Image Inpainting techniques for void filling in glaciological remote sensing products

Thorsten Seehaus1, Bänsch Eberhard2, McNabb Robert3, and Braun Matthias1
Thorsten Seehaus et al.
  • 1Institute of Geography, Friedrich-Alexander University Erlangen-Nürnberg, Germany
  • 2Department of Mathematics, Friedrich-Alexander University Erlangen-Nürnberg, Germany
  • 3Department of Geosciences, University of Oslo, Norway

Remote sensing offers the possibility to efficiently monitor glacier changes on large scales and in remote regions. Glacier surface elevation changes and surface velocities can be derived automatically from satellite acquisitions and provide information on the evaluation of glacier dynamics and mass balance. However, the obtained data sets are often affected by voids due to various issues depending on the imaging technique (SAR, optical). Those missing data on the one hand lead to uncertainties in the quantification of glacier changes, on the other hand can limit the assimilation of the data sets in glacier models.

Inpainting techniques were developed to remove distortions from photographs or for retouch purposes. In this study, suitable Inpainting techniques are applied on glaciological remote sensing products and evaluated in comparison with previous attempts.

For Glacier Bay Alaska, a nearly complete coverage of a glacier area of ~6000 km² by surface elevation change information exists. Artificial voids were generated and filled by using different Inpainting techniques and parameter. The inpainted data sets are evaluated in comparison to the original data set.

How to cite: Seehaus, T., Eberhard, B., Robert, M., and Matthias, B.: Image Inpainting techniques for void filling in glaciological remote sensing products, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2896,, 2020.

Display materials

Display file

Comments on the display material

AC: Author Comment | CC: Community Comment | Report abuse

Display material version 1 – uploaded on 27 Apr 2020, no comments