Accuracy of Surface Water Maps Derived from Radar Satellite Imagery Compared to Multispectral Satellite Imagery DOI Creative Commons
Katelyn Kirby, Colin D. Rennie,

Riley Poot

et al.

Canadian Journal of Remote Sensing, Journal Year: 2024, Volume and Issue: 50(1)

Published: Dec. 5, 2024

Optical and radar remote sensing are both used to map surface water features. Since use cases range considerably in the literature between applications, a direct comparison is warranted assess how well each perform wide of geographic settings using classification methods. Thus, maps generated from Sentinel-1 Synthetic Aperture Radar (S1SAR) Sentinel-2 Multispectral Instrument (S2MSI) imagery were compared across four machine learning techniques eight diverse image areas Canada. Additionally, polarizations multispectral bands varied understand their effect. The results validated high resolution satellite imagery, analysis variance was calculated. S2MSI consistently produced higher accuracy S1SAR. Contrary previous understanding, cross-polarization did not produce significantly more accurate than like-polarization, same true for dual single polarization. introduction an additional band improved significantly. In flooded conditions, polarization best results, detection ice, results. These findings will increase quality efficient generation resource management, climate change impact studies, other disciplines.

Language: Английский

Accuracy of Surface Water Maps Derived from Radar Satellite Imagery Compared to Multispectral Satellite Imagery DOI Creative Commons
Katelyn Kirby, Colin D. Rennie,

Riley Poot

et al.

Canadian Journal of Remote Sensing, Journal Year: 2024, Volume and Issue: 50(1)

Published: Dec. 5, 2024

Optical and radar remote sensing are both used to map surface water features. Since use cases range considerably in the literature between applications, a direct comparison is warranted assess how well each perform wide of geographic settings using classification methods. Thus, maps generated from Sentinel-1 Synthetic Aperture Radar (S1SAR) Sentinel-2 Multispectral Instrument (S2MSI) imagery were compared across four machine learning techniques eight diverse image areas Canada. Additionally, polarizations multispectral bands varied understand their effect. The results validated high resolution satellite imagery, analysis variance was calculated. S2MSI consistently produced higher accuracy S1SAR. Contrary previous understanding, cross-polarization did not produce significantly more accurate than like-polarization, same true for dual single polarization. introduction an additional band improved significantly. In flooded conditions, polarization best results, detection ice, results. These findings will increase quality efficient generation resource management, climate change impact studies, other disciplines.

Language: Английский

Citations

0