SUREHYP: An Open Source Python Package for Preprocessing Hyperion Radiance Data and Retrieving Surface Reflectance

Abstract

Surface reflectance is an essential product from remote sensing Earth observations critical for a wide variety of applications, including consistent land cover mapping and change, and estimation of vegetation attributes. From 2000 to 2017 the Earth Observing-1 Hyperion instrument acquired the first satellite based hyperspectral image archive from space resulting in over 83,138 publicly available images. Hyperion imagery however requires significant preprocessing to derive surface reflectance. SUREHYP is a Python package designed to process batches of Hyperion images, bringing together a number of published algorithms and methods to correct at sensor radiance and derive surface reflectance. In this paper, we present the SUREHYP workflow and demonstrate its application on Hyperion imagery. Results indicate SUREHYP produces flat terrain surface reflectance results comparable to commercially available software, with reflectance values for the whole spectral range almost entirely within 10% of the software’s over a reference target, yet it is publicly available and open source, allowing the exploitation of this valuable hyperspectral archive on a global scale.

Mapping canopy traits over Québec using airborne and spaceborne imaging spectroscopy

“…Due to poor calibration of the pushbroom sensor, Hyperion images contain artefacts (low SNR, stripes, spectral smile...) that have to be accounted for before further processing. In order to obtain surface reflectance, Hyperion images were preprocessed using SUREHYP 71 . SUREHYP is a Python package bringing together multiple methods for destriping, desmiling, and performing atmospheric correction so as to facilitate the processing of a large number of Hyperion images.…”