In situ measurements for the validation of Sentinel-2 Data

Abstract

Validation of remotely-sensed data is crucial for the massive adoption of applications fostering remote sensing data. In the framework of the Copernicus program, the Sentinel-2 data are an invaluable source of information for agriculture. In this abstract we perform two types of in situ measurements for the validation of Sentinel-2 data: (i) by using an FieldScout NDVI meter – which is a portable device capable of instantly measuring the Normalized Difference Vegetation Index widely used for assessing the status of vegetation, and (ii) by using hyperspectral images acquired with the SPECIM IQ – which is a portable camera within the VISNIR (visible and near infra-red) domain (400-1000 nm). From the hyperspectral images acquired in situ we computed the entropy for each spectral band corresponding the Sentinel-2 multi-spectral instrument (MSI), as entropy is often use for the evaluation of grassland quality. We show experimental results from two typers of agricultural crops – common spring wheat and grassland. We notice that the in situ NDVI measurements usually exhibit larger values compared to the NDVI values computed based on the Sentinel-2 data, while the entropy values are usually larger for the satellite data.