Tropical cyclone pattern recognition for intensity and forecasting analysis from satellite imagery

Systematic procedures for both analysis and forecasting of tropical cyclone intensity have been developed over the years. These procedures were designed to improve both the reliability and the consistency of intensity estimates made from satellite imagery. Although procedures have been used and tested under operational conditions at the centers responsible for tropical storm surveillance, some difficulties and complexity are still there. Cloud features are analysed to find out the T-number of the tropical cyclone patterns. This can be solved by reinforcement learning for adaptive tropical cyclone patterns segmentation and feature extraction. Tropical cyclone recognition is a multilevel process requiring a sequence of algorithms at low, intermediate, and high levels. Generally such systems are open loop with no feedback between levels and assuring their robustness is key challenge in computer vision and patterns recognition research. A robust closed-loop system based tropical cyclone forecast method based on reinforcement learning is introduced.