The identification of pseudo-invariant targets using ground field spectroscopy measurements intended for the removal of atmospheric effects from satellite imagery: a case study of the Limassol area in Cyprus

Abstract

This study proposes five types of commonly located surfaces that were evaluated to be determined for use as pseudo-invariant targets (PITs) in order to apply the so-called ‘empirical-line’ atmospheric correction method. Spectroradiometric measurements were taken over wet and dry conditions to obtain the spectral signatures of the targets. From the acquired in situ spectroradiometric campaign, it was found that the proposed commonly found PITs (sandy, concrete and asphalt) are suitable non-variant targets. An accuracy assessment of the empirical-line atmospheric correction method using the five PITs was performed using other calibration targets and aerosol optical thickness (AOT) values from MICROTOPS hand-held sun photometers acquired simultaneously with the satellite overpass. The radiative transfer equation was used to determine AOT levels, using the reflectance values derived from the empirical-line method, in order to conduct an accuracy assessment with in situ AOT measurements. It was also shown that precipitation conditions should be considered when using non-variant targets in atmospheric correction methods.

Acknowledgements

The authors would like to acknowledge the Cyprus University of Technology for the internal funding of this work. The work presented in this study is part of Kyriacos Themistocleous's PhD thesis entitled ‘Improving atmospheric correction methods for aerosol optical thickness retrieval supported by in-situ observations and GIS analysis’.