Angular Dependency of Hyperspectral Measurements over Wheat Characterized by a Novel UAV Based Goniometer

Remote Sensing

Gunter Menz¹  Georg Bareth²  Helge Aasen²  Luis Alonso³  Uwe Rascher⁴  Andreas Burkart⁴ 

[1] Department of Geography, Remote Sensing Research Group, University of Bonn,53001 Bonn, Germany;Department of Geoscience, Working Group GIS and Remote Sensing, University of Cologne, 50923 Cologne, Germany;Laboratory of image processing, University of Valencia, 46980 Paterna, Spain;Research Center Jülich, Institute of Bio- and Geosciences, IBG-2: Plant Sciences,52428 Jülich, Germany;

关键词: hyperspectral;    unmanned aerial vehicle (UAV);    vegetation;    bidirectional reflectance distribution function (BRDF);    goniometer;    vegetation indices;   

DOI  :  10.3390/rs70100725

来源: DOAJ

【 摘 要 】

In this study we present a hyperspectral flying goniometer system, based on a rotary-wing unmanned aerial vehicle (UAV) equipped with a spectrometer mounted on an active gimbal. We show that this approach may be used to collect multiangular hyperspectral data over vegetated environments. The pointing and positioning accuracy are assessed using structure from motion and vary from σ = 1° to 8° in pointing and σ = 0.7 to 0.8 m in positioning. We use a wheat dataset to investigate the influence of angular effects on the NDVI, TCARI and REIP vegetation indices. Angular effects caused significant variations on the indices: NDVI = 0.83–0.95; TCARI = 0.04–0.116; REIP = 729–735 nm. Our analysis highlights the necessity to consider angular effects in optical sensors when observing vegetation. We compare the measurements of the UAV goniometer to the angular modules of the SCOPE radiative transfer model. Model and measurements are in high accordance(r2 = 0.88) in the infrared region at angles close to nadir; in contrast the comparison show discrepancies at low tilt angles (r2 = 0.25). This study demonstrates that the UAV goniometer is a promising approach for the fast and flexible assessment of angular effects.

【 授权许可】

Unknown