Texture Based Vector Field Visualization

Vectorfieldvisualizationisoneofthedisciplines withthemostchallengein scientific visualization due to its extension application, large data sets andso on. Meanwhile, texture based vector field visualization, which is able todisplay detailed vector field information continuously, is the most broadlyappliedvectorfieldvisualizationapproach.On the basis of a large number of theories, including computer graphics,digital image processing, calculus, computational fluid dynamics andnumerical methods, this dissertation investigates the previous texture basedvectorfieldvisualizationapproaches,suchas Line Integral Convolution(LIC)and its enhanced methods. This dissertation proposes a fast integralconvolution withSobol sequence (Sobol-FLIC), which can not only show thevector field accurately and intuitionally, but also accelerate the computationprocedure.This approach utilizes the Sobol sequence to achieve better coverage ofthe image with less streamlines. As a result, it enhances the efficiency of thecomputation. Meanwhile, it can generate sparse texture and dense texturesimultaneously. So the sparse texture, which is able to display the fieldstructure as well as improve the efficiency, can be used directly when thedetailofthevectorfieldisnotimportant.Alsothepropertyoftheresultimageis enhanced by using the Double LIC method. It is proven that the newmethod is more efficient and has better image coverage with fixed number ofstreamlinesthansomeothermethods(LIC,FLIC)inMatlabenvironment.