Research And Application Of LIC Texture Visualization In The Flow Field

Computer graphics and image processing technology have been used in flow visualization technology to explore the characteristics of the flow field. It can provide a powerful tool to study flow in natural and make decision of major engineering projects. Particularly the texture visualization technology is widely used as it can describe the characters information of flow field with shapes and colors. But there are many disadvantages such as huge computation, low rendering efficiency and so on. When extended to 3D flow field, the problem of occlusion and confusion is exited between texture slices. To solve these problems and shortcomings, many contributions had been made in this thesis as follows:In the case of 2D LIC texture visualization, when the uneven distribution of flow field property values occurs, characteristics of the flow field in details cannot be described with the overall distribution of color in the rende ring images produced by traditional linear color mapping method. In this study, a new color mapping algorithm based on the hierarchical clustering was proposed to effectively solve this problem by overall sampling, adjacency clustering and isometric mapping of color gamut. The testing results on simulative wind field dataset and the Isabel Hurricane sectional dataset have suggested that the rendering images were vivid and delicate. And our method can effectively relieve the phenomenon of color clustering.To solve the problem of occlusion and low rendering efficiency with traditional Volume LIC algorithm, the Volume LIC rendering technology based on the sparse noise accelerated by GPU was proposed in this thesis. In order to reduce the phenomenon of space block, a sparse noise created by the sparse noise generation algorithm based on the Hilbert space filling curve was used to replace the original dense noise. Meanwhile, the hardware accelerating strategy based on GPU+GLSL was designed to optimize the complex ity of the algorithm and improve the efficiency of the algorithm. Furthermore, the testing results with Tornado dataset and Isabel Hurricane simulation dataset indicated that time efficiency and rendering effect was obvious improved in our algorithm compar ed with the traditional Volume LIC.A three-dimensional flow field texture visualization prototype system was designed in this paper. And the above algorithms were integrated into this system. Additionally, there also were some other functions like data ch aracteristics topology demonstration, two-dimensional flow field texture visualization, and three-dimensional texture visualization, etc. Multi-datasets have been used to test the function and interaction of this system. Experimental results showed that th e system has ran fluently and provided an excellent visualization effect.