Share Email Print
cover

Proceedings Paper

On accuracy of radiometric calibration of hyperspectral visible/NIR satellite remote sensing instruments above natural surfaces
Author(s): Alexander Borovski; Oleg Postylyakov
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

To provide accurate data the regular on-board absolute radiometric calibration of a satellite hyperspectral instrument is required. Together with the internal calibration the external calibration using comparison of radiance measurements above special ground test sites and calculated radiances is performed. The top of the atmosphere radiances are calculated using a radiative transfer model basing on atmospheric and surface characteristics measured at the test sites. Spatial pixel of hyperspectral instruments is tens of meters, therefore it is difficult to creation with large enough fields with artificial reflective surface. The paper presents preliminary results of the theoretical analysis of the errors of a satellite hyperspectral instrument radiometric calibration using grass and snow. Results for test sites located at 200 m.a.s.l. and 2000 m.a.s.l. with the atmospheric composition and surface reflectance measurements are compared. The analysis is performed for an instrument with the spectral resolution of 1-8 nm which is typical for special regime of payload GSA of Russian satellite Resurs-P. The errors related with the atmospheric composition and albedo measurement errors and scenarios of the aerosol vertical distribution were theoretically examined. The error is less than 6% in all the cases at all the wavelengths between 400 nm and 1000 nm with the exception of the absorption bands of water vapor. In the absorption bands of water vapor about 720 nm and 820 nm the errors reach 12% at the downcountry site. In the absorption band of 950 nm the errors reach 15% in mountains and 35% in downcountry. In our error analysis we used 4% error of surface albedo. We can expect that reassessment of the albedo error of natural surfaces will directly influence to the estimation of the total errors for the both grass end snow except the spectral range of 400-650 nm where grass has low albedo.

Paper Details

Date Published: 29 November 2016
PDF: 9 pages
Proc. SPIE 10035, 22nd International Symposium on Atmospheric and Ocean Optics: Atmospheric Physics, 1003513 (29 November 2016); doi: 10.1117/12.2248632
Show Author Affiliations
Alexander Borovski, A.M. Obukhov Institute of Atmospheric Physics (Russian Federation)
Oleg Postylyakov, A.M. Obukhov Institute of Atmospheric Physics (Russian Federation)


Published in SPIE Proceedings Vol. 10035:
22nd International Symposium on Atmospheric and Ocean Optics: Atmospheric Physics
Gennadii G. Matvienko; Oleg A. Romanovskii, Editor(s)

© SPIE. Terms of Use
Back to Top