Harmonization of hyperspectral and multispectral data sets

Kurzfassung

Many applications of earth observation missions rely on time series data. Combination of data from different missions makes the time series denser and thus has the potential to provide more information. However, data from different missions have to be harmonized to be analysed together in one time series. The Sen2Like project started to harmonize Sentinel-2 and Landsat-8 data and included an investigation how to add data from hyperspectral missions. The harmonization process includes a consistent cloud screening and atmospheric correction, spectral adjustment, BRDF adjustments and re-gridding. The presentation will give some recommendations on harmonisation of data formats and products to make easier the data fusion process. Multispectral earth observation data from several missions are already available since decades thus providing an incredible time series for many applications. These time series cannot be replaced by hyperspectral data, but supplemented. Implementation of hyperspectral data into multispectral Sen2Like framework is possible in two different ways regarding cloud screening and atmospheric correction. Either the atmospheric correction is performed on hyperspectral L1C-data or on multispectral L1C data generated from the hyperspectral origin. Results show only small differences between both ways.