Prediction the Direction of Storm Based on the Sub Pixel Registration

Abstract

In the field of natural disaster management, storm detection, forecasting the weather condition and change detection are occupied as essential role. The storm direction detection is provided to manage and to prevent the possible attacked region of the land. The storm regions are extracted by the feature extraction method based on Morphological processing and the color segmentation approach based on the intensity transformation and color spaces. The movement of the interested cloud region are estimated by the modified sub pixel registration approach. The direction of the storm can be tracked from these computations. Especially, the attacked region of Myanmar Land is forecasted from multi-date satellite images. The digital image correlation method is developed based on the subpixel accuracy method. According to the experiment and its result, the proposed method work effectively.