A comparison of methods for the retrieval of surface reflectance factor from multitemporal SPOT HRV, HRVIR, and HRG multispectral satellite imagery

作者：Clark Barnaby^*; Suomalainen Juha; Pellikka Petri

来源：Canadian Journal of Remote Sensing, 2010, 36(4): 397-411.

Absolute atmospheric correction for retrieval of the surface reflectance factor (rho(s)) is a prerequisite for quantitative remote sensing utilizing multispectral optical satellite data. Areas where changes in land cover are occurring with greatest significance also often lack detailed meteorological data that allow full use of radiative transfer models (RTMs) for reflectance factor retrieval (RFR). This study assessed RFR accuracy of techniques applicable in such circumstances. The historical empirical line method (HELM), four image-based dark-object subtraction (DOS) methods, and the 6S RTM were applied to SPOT imagery datasets covering both the Taita Hills application site in southeast Kenya and the Helsinki metropolitan region control site in Finland. HELM was the only approach that achieved RFR in the visible and near-infrared bands with a RMSE of <0.02 rho(s) and overall relative accuracy of <10%. Further, HELM shortwave infrared (SWIR) performance was significantly better than that of the other techniques and the partially corrected at-satellite reflectance (rho(SAT)). Although better than the DOS methods, application of 6S using standard atmosphere and aerosol models did not meet the desired RFR accuracy, even with the utilization of horizontal visibility meteorological data.

• 出版日期2010-8