VemoSat

The map shows (left) a Germany mosaic from Sentinel-2 data (S2)* with a native image resolution of 10 x 10 meters and Super Resolution images (right) with an image resolution of 2.5 x 2.5 meters, which are created by sharpening the Germany mosaic from Sentinel-2 data.
The routes of high (110 kV) and extra-high voltage lines (220/380 kV) are extracted and processed from OpenStreetMap**.

Short description of the project

The core of the Vegetation Monitoring for Energy Infrastructures on a Satellite Basis (VemoSat) project is monitoring vegetation changes due to growth and impact events such as tree breakage or tree throw using Earth Observation (EO) data along vulnerable infrastructure like high voltage power lines.
The goal is to test already freely available Earth Observation (EO) data like Sentinel-2 in the area of vegetation monitoring and to improve their detection accuracy.

Methodology Description

Common methods to improve spatial resolution in remote sensing include the use of interpolation filters (e.g., bicubic) or pan-sharpening techniques when a higher resolution panchromatic image is available. In the project, we implement and train a model based on the Enhanced Super-Resolution Generative Adversarial Network (REAL-ESRGAN) (Wang et al. 2021)***. Image pairs from digital orthophotos**** (DOP) and Sentinel-2 data are used to produce a super-resolved multispectral Sentinel-2 output with a scaling factor of 4.

* © European Union, contains Copernicus Sentinel-2 data [2021], processed by the Federal Agency for Cartography and Geodesy (BKG)
** © OpenStreetMap
*** Wang, Xintao; Xie, Liangbin; Dong, Chao; Shan, Ying (2021): Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data. Online hier verfügbar.
**** © GeoBasis-DE / BKG (2018)

Logo Kognitive Energiesysteme

Fraunhofer IEE | VemoSat 🇩🇪 🇬🇧

VemoSat

Short description of the project

Methodology Description

Hilfe

Weitere Forschungsaktivitäten

Kontakt