Deep Neural Network Regression for Normalized Digital Surface Model Generation with Sentinel-2 Imagery

Deep Neural Network Regression for Normalized Digital Surface Model Generation with Sentinel-2 Imagery

November 15, 2023

Our EAGLE student Konstantin Müller published together with our chairholder of the Deparment of Global Urbanization and Remote Sensing, Hannes Taubenboeck an article about DL for surface model generation. The article explores methods to extract high-resolution normalized digital surface models (nDSMs) from low-resolution Sentinel-2 data, enabling the creation of large-scale models. Leveraging the open access and global coverage of Sentinel 2, the study employs deep learning models, based on the U-Net architecture, with tailored multiscale encoders and conformed self-attention to achieve a mean height error of approximately 2m.

from the abstract: In recent history, normalized digital surface models (nDSMs) have been constantly gaining importance as a means to solve large-scale geographic problems. High-resolution surface models are precious, as they can provide detailed information for a specific area. However, measurements with a high resolution are time consuming and costly. Only a few approaches exist to create high-resolution nDSMs for extensive areas. This article explores approaches to extract high-resolution nDSMs from lowresolution Sentinel-2 data, allowing us to derive large-scale models. We thereby utilize the advantages of Sentinel 2 being open access, having global coverage, and providing steady updates through a high repetition rate. Several deep learning models are trained to overcome the gap in producing high-resolution surface maps from low-resolution input data. With U-Net as a base architecture, we extend the capabilities of our model by integrating tailored multiscale encoders with differently sized kernels in the convolution as well as conformed self-attention inside the skip connection gates. Using pixelwise regression, our U-Net base models can achieve a mean height error of approximately 2 m. Moreover, through our enhancements to the model architecture, we reduce the model error by more than 7%.

Link: https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=10189905

you may also like:

RIESGOS project completion

RIESGOS project completion

Today, the six-year international project RIESGOS, led by DLR-DFD, funded by the Federal Ministry of Education and Research (BMBF) came to a successful completion. Dr. Elisabeth Schoepfer from DLR-DFD – one of our guest lecturers here at EORC – led this...

Experiencing rural Würzburg – NetCDA on a cultural exchange

Experiencing rural Würzburg – NetCDA on a cultural exchange

This week we went with our guest graduate students from West Africa on a small trip to some rural areas of Würzburg. Cultural exchange is an important component for our students – for some students it is their first visit to Germany, often also to Europe. Thus,...

Kickoff Meeting of the FluBig Project

Kickoff Meeting of the FluBig Project

On Tuesday, 27th of February, the kickoff meeting of the DFG funded projected FluBig took place at the EORC in Würzburg. Jointly with colleagues from the Karlsruhe Institute of Technology, Catholic University Eichstätt-Ingolstadt and Earth Observation Research...

University press covered our high alpine activity

University press covered our high alpine activity

The university press featured our activity in the Alps on the research station "Schneefernerhaus" close to Zugspitze. Together with other research groups of our university we aim to conduct various field data acquisitions e.g. in collaboration with the biology...

High alpine and snow training

High alpine and snow training

In collaboration with the German Alpine Club (DAV), Laura, one of our EAGLE students and Clara, a student assistant at our institute, recently participated in avalanche rescue training aimed at equipping individuals with essential skills and knowledge to respond...

Privacy Policy

Lehrstuhl für Fernerkundung & Lehrstuhl für Urbane Fernerkundung

Erdbeobachtung an der Universität Würzburg