Research On Visualization Technology Of Meteorological Vector Field Data

Visualization in Scientific Computing is a theory and technology that convert data into graphics or images and then display them on the screen with the use of computer graphics and image processing technology.Vector field visualization is an important subject in the research of scientific computing visualization,and it has been applied in weather forecast,hydrodynamics,aeronautics and Astronautics,turbine design and so on.In the field of meteorology,the visualization of vector field data and the study of meteorological regularity can prevent disasters and reduce casualties and property losses.This thesis systematically introduces various techniques used in visualization of meteorological vector field data.And this thesis mainly researches streamline visualization and texture-based visualization.For the streamline visualization method,this thesis mainly considers how to select seed points of the streamline and how to attach color to the streamline,and starts research with those problems.For how to select streamline seed points,Sobol sequence is used to determine the location of seed points to get multiple streamlines with relatively uniform distribution.In order to show more information of vector field data,color is attached to the streamline.For the texture-based vector field visualization method,this thesis mainly considers the problem that how to merge the gray texture after line integral convolution and the color value mapped by the vector field data,and improves the visualization process,and analyses visualization effects.The main innovations and the major tasks completed in this thesis as follows:(1)For the problem of how to present the information of vector field strength on streamlines,a novel streamline coloring method is proposed,which adaptively chooses the mapping method according to the distribution features of vector field strength.Firstly,the skewness coefficient of the input vector field strength data are calculated,then compared with the threshold to determine the color mapping method.Finally,the streamlines are colored when they are generated,and the streamline color can more accurately express the vector field strength information.(2)For the problem of how to present vector field direction information through the texture,a color enhancement algorithm of LIC texture with a legend is proposed.The values of gray texture after two line integral convolutions and the value mapped by vector field data in HSV color space are enhanced respectively,and merge them to get the color texture.Finally,the visualization results are displayed with a legend to solve the ambiguity of vector direction in texture images.(3)A static visualization method is adopted for vector field data,including point icon method and clustering method.SOM clustering method is used to cluster vector field data on vector direction.In the step of color mapping,the similarity with the winning neuron is added,which makes the internal color of each cluster change in light and shade,and the direction information of vector field can be expressed better.(4)A dynamic visualization method is adopted for vector field data,including the dynamic texture noise method and the particle convection in 2D or 3D space.The dynamic noise texture method merges and updates the noise along the vector direction in each frame.And combining with the color mapping,the vector texture,vector direction and vector size are dynamically expressed in a way like particles move.Finally,a visual plugin is constructed based on the implemented algorithm,which is integrated into the three-dimensional platform to show the final experimental results.