| Popis: |
Visible cloud image of high resolution satellite (HJ-1) can clearly show the details of the eye of the typhoon cloud system. In the paper, a new artificial intelligence algorithm about the eye of typhoon center position would be proposed in accordance with this feature. Firstly, the image segmentation technique of entropy-based thresholding is used to extracted typhoon cloud features, then used the methods of morphology opening operation and seed the filling algorithm to obtain the contour of the typhoon eye. In addition, we applied the gravity calculation method to locate the center of the typhoon. Comparison the typhoon location data which obtained by the new method to the Central Meteorological Observatory, the experimental results showed that the deviation is smaller, and this can be proved that the algorithm is effective. |
