Mesocyclone Recognition Based On Doppler Radar Image

The work of this paper belongs to the project"The research of automatic hailstone detection system based on Radar".Of all natural disasters, hailstone is most dangerous, which threats to life and property. The hailstorm arises suddenly and locally, moves fast, and dissipates very quickly, so it is difficult to forecast. Actually a powerful support is supplied by Doppler radar. Nowadays the research of hailstone forecasting based on Doppler radar image is at its developing stage. Therefore, this project enjoys a highly research value.In the past, almost all the research is based on the data. Different from that, this paper researches the image of Doppler radar with pattern recognition and artificial intelligent methods from the view of image processing. Together with many years'experiences of experts, a good method is put forward.Three parts are included in the paper as follows:(1) Two images with different elevation angle are synthesized. As a result, some noises are filtered.(2) Firstly, the scope of the image is limited clipping. Secondly, the pixels which means low velocity and aliasing velocity are dealt with by segmentation. Then the reflectivity image removes the pixels out of the range of 50dBZ.(3) If the number of the area connected is 2, eigenvalue extracted from the image will be used to locate the position of mesocyclone. And at the same time, mesocyclone is classified.Theoretical analysis and experimental results demonstrate that, methods used in this paper are effective for mesocyclone recognition based on Doppler radar, and can be applied to the practical situations. This research will be a solid base for the further investigations on mesocyclone recognition based on Doppler radar.