Fusion of Satellite and Drone Images for Identifying Water Bodies

• Ariane Paula Barros CPQD http://orcid.org/0009-0009-7420-4798
• Elizângela Jussara dos Santos CPQD
• Fernanda Kristen da Silva Pedro CPQD https://orcid.org/0009-0003-7336-7786
• Dimas Augusto Mendes Lemes PUC-Campinas https://orcid.org/0000-0001-5189-4129
• José Picolo PUC-Campinas https://orcid.org/0009-0002-0214-4932
• Guilherme Ribeiro Sales CPQD https://orcid.org/0009-0004-7994-7973
• Valentino Corso CPQD https://orcid.org/0009-0004-5253-5151
• Cides Semprebom Bezerra CPQD

Abstract

Mapping areas with water bodies is crucial for monitoring water systems and resource management, as well as for guiding urban expansion and land-use planning, thus contributing to sustainable resource management, flood prevention, and the conservation of aquatic ecosystems. While satellite images offer broad territorial coverage, drones provide high spatial resolution of the Earth's surface. In this research, the Pulse Coupled Neural Network (PCNN) [1] image fusion method and the GDAL Pansharpening method, the latter being available in QGIS, were compared to determine which is more effective in merging images to highlight water surface areas in a case study, combining the spectral coverage of satellites with the precision of drones. The implemented PCNN model was able to overcome QGIS Pansharpening method, which is based on Brovey algorithm, obtaining better metrics. This study of drone and satellite image fusion has shown to be a promising approach to overcome the limitations of these technologies when used individually.

Keywords: Image Fusion, Remote Sensing, Drone, Riverine Systems, High Resolution