Downburst Recognition Based on a New Watershed Algorithm

摘要

This paper studies the automatic identification of downburst based on the images of Doppler radar with pattern recognition methods from the view of image processing. Together with the characteristics of extreme region corresponding to the downburst in Doppler radar images and the experiences of experts, a new watershed algorithm is proposed based on the regional properties bound. Experimental results and theoretical analysis demonstrate that methods used in this paper are effective for downburst recognition, and can be applied to the practical situations.