Knowledge based system for weather information processing and forecasting

Using physical observations from the satellite imagery and the meteorological information a knowledge based weather information processing and forecasting system (KB/WIS) is designed the characteristic features extracted from the weather images are used to represent various weather patterns. The system is statistically trained to characterize, interpret, and forecast weather patterns. The system proposed is a composite of five components, namely: image acquisition, image preprocessing and enhancement, feature extraction and selection, weather knowledge base, and weather INference engine (WINE). Complete architecture of the KB/WIS system including its application and algorithms is described.