Motion Analysis Based On Semi-lagrangian In Radar Images

The doppler weather radar is the main detection tool for short-time severeweather.Using the combination of reflectivity factors of radar image, this paperstudies the application of the semi-Lagrangian methods on the motion vector fieldanalysis and the forecast of radar images. The main research content of this article isdivided into the following two parts:The paper uses L-K optical flow method to calculate the radar echo motionvector.In the using L-K optical flow method, it gives the more reasonable parametersneeded in the calculation process. And it uses the block processing methods whichimproves the calculation rate.Finally,it summarizes the detailed steps of the algorithm,and through the analysis of the experiments it shows that the method could calculatethe radar echo motion vector field effectively.The paper studies the application of semi-Lagrangian method on the forecast ofradar echo images.Semi-Lagrangian extrapolation method extrapolates along thetrajectory of radar echo motion vector on the basis of radar image vector field.It canrealize curve extrapolation which has important significance for typhoon and the otherrotary radar image extrapolation forecasting.To achieve the goal of forecast it uses theafter extrapolation method in the process of calculation,that is,from the end grid alongthe proton trajectory it extrapolates to the place of the starting point, calculating thecoordinates and all kinds of values of the middle point and the starting point. Valuesare calculated respectively by using bilinear interpolation and cubicinterpolation.Then through a lot of experiments,it compared and analyzed the forecastresults of two and three time layer semi-Lagrangian extrapolation method. Finally,itchoosed the two time layer semi-Lagrangian extrapolation method for the stormnowcasting prediction.