A Matrix-based Satellite Cloud Tracking Algorithm

Solar power generation is heavily reliant on the availability of direct solar irradiation. Intermittent appearances of clouds over a solar plant introduce high variability to solar power output and result in high forecast errors. This warrants the study of different cloud patterns. In this paper, we present a novel approach to track different cloud patterns throughout their lifetimes. We define a matrix, called the Association Matrix, to describe the relationship between different clusters and put forth 7 different conditions for identifying the different cloud phenomena. We use the cloud mask data from the INSAT-3D satellite for our analysis and comprehensively describe the operation of the proposed algorithm in this paper.