Research and Analysis on the Angles Normalization of the Domestic GF-1 Satellite Based on MODIS BRDF Products

Abstract

With the widespread application of multi-source remote sensing data, there are higher requirements for the consistency of data interpretation of different sensors. In view of the fact that the wide space coverage of the wide-format multispectral sensor carried by the domestic GF-1 satellite has a large observation field of view, a normalized model of the angle of the domestic high-resolution satellite is proposed to eliminate the differences caused by the observation geometry between the wide-width multispectral sensors of the domestic high-resolution satellite. Landsat-8 images are used as auxiliary data for orthography correction, and cross radiometric calibration method and atmospheric correction of 6S model are used for radiometric correction. Combined with MODIS low-resolution BRDF product data, the linear semi-empirical kernel-driven model was used to conduct BRDF correction for reflectivity data, and the observation angle is normalized to zenzenical direction. The reflectivity data of the Landsat-8 product is then used to compare with the effects. The results show that the normalized images have better reflectance information, and the rates of change in the first band of each sensor are 12.7%, 5.33%, 3.8%, and 5.07%, respectively, which confirms the applicability and feasibility of this angle normalization method.