CLOUD DETECTION FOR MERIS MULTISPECTRAL IMAGES

摘要

Two of the key features of the MEdium Resolution Imaging Spectrometer (MERIS) instrument on board the ESA Envisat environmental satellite are its temporal resolution of three days and its spatial coverage. The amount of images acquired over the globe every day makes inevitable that many of these images present cloud covers. From an operational point of view, an automatic and accurate method for cloud detection in MERIS scenes is a key issue since, with no accurate cloud masking, undetected clouds are the most significant source of error for true ground reflectance, affecting a wide range of remote sensing applications. By masking only the image areas affected by cloud covers, the whole image is not necessarily discarded. The main objective of this work is to develop and validate a method for cloud detection using self-contained information provided by MERIS Full Resolutiuon Level 1b products and a Digital Elevation Model (DEM). The method must be capable of: detecting clouds accurately providing probability and cloud abundance rather than flags; and better describing detected clouds (cloud abundance, type, height, subpixel coverage) in order to include this information in radiation models.