Atmospheric Correction of HJ-1A CCD Images for an Agricultural Area Using the 6S Model, and Its Influence on Five Vegetation Indices

摘要

Vegetation indices calculated from remote-sensing images are commonly used in agricultural applications. To decrease the influence of the atmosphere on these indices, atmospheric correction is essential. The HJ-1 CCD produces images with high temporal and spatial resolution, and these images therefore have great potential in agricultural applications. To improve the quality of these images, we performed atmospheric correction for an agricultural area in northern China using the 6S model to determine the influence of these corrections on the accuracy of the vegetation indices (compared with ground-level measurements in the field). The 6S model simulates the radiative transfer process between the surface, atmosphere, and remote sensor, thereby acquiring the apparent radiance and apparent reflectance. The surface reflectance is then retrieved and the vegetation indices are computed. We validated the atmospheric correction by comparing the spectral reflectance measured in the field with the spectral reflectance after atmospheric correction and determined the value of NDVI and four other indices before and after the correction. Experiments indicate that the corrected reflectance computed using this method has good reliability compared with reflectance values measured in the field; the mean difference between the two values was less than 0.059.