Research On Feature Extraction Of Flow Field Based On Information Entropy And Visualization

Visualization is a kind of technology to converting data into graphics and realizing interact processing,which is an important means to showing the characteristics of data and the data correlation.Vector field data as a representative in the field of scientific visualization,how to extract the feature information and display t he inner laws of data intuitively which are research focus in the visualization.This paper from two aspect: visual rendering and visual interaction,The mathod of feature extract is based on the theory of entropy-information.The mathod of interaction base on multiresolution.Firstly,in view of the problem of missing detection of critical points,we put forward a kind of new critical point detection method based on local maximum entropy value.Calculating the local maximum entropy on the entropy field,and confirming the candidate point,then according to the Linear interpolation principle to determine whether a candidate point of local area is zero,finally to according to the isoline principle to determine the location of the critical point,which can effectively avoid the problem of feature omissions.Secondly,in order to solve the problem that it is difficult to define the local feature area accurately.We implove the traditional method by put forward a precise definition algorithm based on the change rule around feature points.Centering on the entropy that corresponds to the critical point,and calculating entropy rate,extending outward,when the rate of change is zero reached the edge of the feature area.The local area can be defined accurately based on this method.Thirdly,the traditional vector field visualization based on streamline is with single streamline density,which makes the visualization effict not clear on the amplification of local area.we designed a kind of vector field interactive visualization method based on multi-resolution.According to the difference of viewpoints select the corresponding resolution of the seed point layout density,whether global or local display are able to use appropriate density of streamline visualization.Finally,the traditional method based on a Vector field data visualizationg schema which completes the above mentioned methods is designed,conducting the experiment with the ocean flow field data.The experiment proves the validity of those methods put forward by this paper.