RETRIEVING TOPSOIL PROPERTIES THROUGH MULTIPLATFORM AND MULTI -HYPER SPECTRAL EO DATA

The assessment of soil variables from multispectral remote imagers is hindered by inadequate spectral resolution, therefore they are mainly used for qualitative assessments. This project represents the initial phase of retrieving topsoil properties with multiplatform and multi-hyper-spectral EO data using machine learning and multi variate regression. The study areas are in the experimental farmin QuzhouCounty, HandanCity, HebeiProvince, China. Measurements of topsoil properties like Soil Organic Matter(%), pH, EffectivePhosphorus(mg/Kg), AvailablePotassium(mg/Kg), and SoilNitrogenContent(g/Kg), were carried out between 2019 and 2020. Satellite data from European Space Agency(ESA)Sentinel-2 and the Italian Space Agency(ASI) Hyperspectral Precursor and Application Mission(PRISMA) were used. To improve the machine learning regression of soil properties the images were co-registered and pre-processed.