Abstract

Spatiotemporal fusion (STF) methods are a paramount solution for generating high spatial and temporal time series, overcoming the limitations of spatial and temporal resolution of satellite data. STF methods typically rely on band-by-band fusion, assuming spectral similarities. However, selecting the optimal band for fusion becomes challenging when multiple narrow bands overlap with the target band, often leading to the use of only one single band. Furthermore, sensor specifications and observation configurations can further compound this challenge, reducing spectral and spatial information. We introduce a new preprocessing step that maximizes the use of spectral information from narrow bands. It minimizes radiometric differences caused by sensor variations in the STF process by considering the spectral response function (SRF). Our method generates adjusted bands that closely match the target band's spectral characteristics, leveraging all available spectral information. We evaluated this strategy at two study sites employing Sentinel 2 and Sentinel 3 data by comparing fused images from adjusted bands and the original bands using three popular STF methods. The results obtained showed that the images fused with the adjusted bands were closer to the target images and achieved better performance, improving the fusion quality compared to the original bands (SAM by 37% and RMSE by 30%). The preprocessing step offers a feasible approach to generate spectral bands that would be captured by the sensors if they had the same spectral characteristics. Importantly, this preprocessing technique is applicable to any STF method.