How Effective Is Super-Resolution to Improve Dense Labelling of Coarse Resolution Imagery?

Coarse resolution remote sensing images, such as LANDSAT and MODIS are easily found in public open repositories and, therefore, are widely used in many studies. But their use for automatic creation of thematic maps is very restrict since most of the deep-based semantic segmentation (a.k.a dense labe...

Full description

Saved in:
Bibliographic Details
Published in2019 32nd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI) pp. 202 - 209
Main Authors Pereira, Matheus B., dos Santos, Jefersson A.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.10.2019
Subjects
Image segmentation
Remote sensing
Satellites
semantic segmentation
Semantics
Spatial resolution
super resolution
Task analysis
Online AccessGet full text

Cover

More Information
Summary:Coarse resolution remote sensing images, such as LANDSAT and MODIS are easily found in public open repositories and, therefore, are widely used in many studies. But their use for automatic creation of thematic maps is very restrict since most of the deep-based semantic segmentation (a.k.a dense labelling) approaches are only suitable for subdecimeter data. In this paper, we design a straightforward framework in order to evaluate the effectiveness of deep-based super-resolution in the semantic segmentation of low-resolution remote sensing images. We carried out an extensive set of experiments on three remote sensing datasets with distinct nature/properties. The results show that super-resolution is effective to improve semantic segmentation performance on low-resolution aerial imagery. It not only outperforms unsupervised interpolation but also achieves semantic segmentation results comparable to high-resolution data.
ISSN:2377-5416
DOI:10.1109/SIBGRAPI.2019.00035