An Approach for Spatial Statistical Modelling Remote Sensing Data of Land Cover by Fusing Data of Different Types

KThe article discusses the application of spatial statistical models in Remote Sensing (RS), which appears to be an important source of spatial data at multiple scales. However, to derive the maximum information, it is necessary to apply appropriate techniques of spatial statistics by de-fining a spatial statistical model. A crucial problem facing us is the fusion of multi-source spatial data of different nature and characteristics. One of these properties is the support of measurement, which unfortunately is little considered in RS. Therefore, a multivariate geostatistical approach is proposed that can fuse data of different types, including soil sample, hyperspectral, multispectral UAV and satellite data, considering the various supports and their impacts on prediction and its uncertainty. The method is applied to soil data after filtering UAV images of an olive grove. The fusion of soil sample data with hyperspectral data measured in the laboratory, UAV and satellite data (Planet and Sentinel 2) improved the prediction accuracy of the soil properties compared with univariate prediction. It is hopeful that advanced spatial statistical techniques will be increasingly applied in RS in the future.