Earticle

The recognition and center position methods of typhoon with high accuracy are essential for forecasting and monitoring of tropical cyclones (TCs) in the meteorological applications. We designed the improved adaptive threshold processing algorithms and segmentation of typhoon cloud system by using two value image connecting region labeling algorithm. It shows potential ability in typhoon cloud feature and the recognition of typhoon cloud system in northern hemisphere. We also designed algorithms of locating typhoon center by mathematical morphology. Experimental results show that the typhoon cloud system can be effectively recognized on could imagery and the center of cloud imagery can also be located by the proposed method. Our analysis indicated that typhoon recognition and center position accuracy based on FY-2 satellites data were high recognition rate and accuracy of the center location, and the methods can be applied to meteorological business systems.