Impact of sun-view geometry on canopy spectral reflectance variability

Abstract

UAS-based remote sensing has been rapidly growing in numerous fields, including agriculture and environmental studies. However, the lack of a practical and reliable quantitative method independent of the experiment condition has limited the translation of remote sensing models to another time and space since the results are barely generalizable. Several error sources severely affect the UAS-based remote sensing data; one of the most important ones is the sun-camera geometry. This study investigates the effect of sun-camera geometry on the reflectance of several specialty crops, including citrus, almond, and grape, quantifies the magnitude of the effect, and introduces a practical approach to account for these effects. We learned that a minimal variation of 2° in view angle causes a significantly different reflectance for similar canopies. Due to the directional solar radiation, the canopy reflectance measured from different view angles in the cameras' Field of View (FOV) could vary more than ± 50 % of the nadir view. Despite widely practiced assumptions in the literature, results show that these variations are not negligible, and the practical use of remote sensing imagery requires compensation for directional effects. We developed a model based on the Laplacian distribution function that estimates the variations in aerial images in two directions: principal plane and cross-plane. The nadir reflectance of each canopy can be calculated from other view angles using the developed model with r2 of up to 0.88 based on the band, crop, and other variables. The 4Sail radiative transfer model was used as a benchmark to evaluate the model’s performance. We leveraged the output of this study to determine the optimum flight time as a function of date, the camera’s FOV, and the site’s location latitude, which will facilitate the correction of the solar angle effect and help avoid unreliable data collections.

Topics & Concepts