Classification of cloudiness from MODIS satellite data using regional statistical models for image texture and physical parameters of cloudiness during periods with snow cover

摘要

Statistical models are presented of the image texture and cloudiness physical characteristics over various natural zones of the Russian Federation during periods of snow cover. These models are based on the determination of the distribution laws and estimation of their parameters, which describe the fluctuations in the values of the cloud characteristics. The results are discussed of a comparative analysis of statistical cloud models for various natural zones, as well as cloud models, averaged over them, over snow-covered territories and a snow-free underlying surface. A description is presented of the cloud classification algorithm based on the application of artificial neural network technology and fuzzy logic methods. The results are presented of recognition of 12 cloud types from MODIS satellite data for various natural zones during seasons with snow cover.