Study On Shape And Formation Extrapolation Algorithm For Cloud Of Storm Based On Doppler Weather Radar

Hailstones, squall line and rainstorm is one of disastrous weathers and it lasts for short time and is destructive .Up to now, forecasters who accord to their artificial experience analyses Doppler radar images conclude prediction of short-term severe meteorological weather.Based on consecutive reflective image of Doppler Radar, this article combines image processing technology and computer software engineering in order to introduce a whole new algorithm of image extrapolation. Some new traits are brought as follows:With the difference of shape and area of cloud clusters, two extrapolation methods are introduced. If the area of cloud clusters is small and its shape is bulk-shaped, the scattering model method is applied to this kind of cloud clusters. If the area of cloud clusters is large and its shape is belt-shaped, corrosion expansion method of Fan-shaped edge is applied to this kind of cloud clusters.According to the trait that reflectivity of severe convective weather decreases gradually from interior area to exterior area, decompose the cloud into many layers. Make every layer be one connected region with one kind of value, then, separate this layer into many blocks according to the algorithm that the irregular shape can be separated into some regular blocks which two extrapolation methods can be applied to through multi-block decomposition algorithm. If the shape of cloud clusters is regular, two methods can be applied to cloud clusters directly. However, the shape of some cloud clusters is not regular, multi-block decomposition must be carried before extrapolation in order to make two algorithms work more effectively.Design and implement synthetic extrapolation program to every block.The algorithm introduced above has been implemented by Visual Studio and OpenCV library and has been tested. Theoretical analysis and experiment shows that the research of this dissertation has made great achievement that automatic detection and prediction of severe convection weather system.