Research On Multi-Dimensional Visualization Technology For Global Marine Reanalysis Product

Nowadays,the complicated system and huge amount of marine scientific data provide strong support for the research on marine environment,marine meteorology,marine forecasting,and even marine military.Among them,the marine environmental element data plays a vital role.How to process and analyze these data is a topic worthy of attention.Combined with visualization technology to achieve this goal,it can more objectively and directly describe the specific information and changing process of the marine environment,and simulate it.Based on the marine environment data visualization requirements in the project of "Global High Resolution Ocean Reanalysis Product Application Demonstration",this paper studies the visualization technology of marine environment element data in multiple dimensions and types,and uses Graphics Processing Unit(GPU)to accelerate the original rendering.Then the algorithms were optimized for the lack of them,and a multi-dimensional visualization system for ocean reanalysis products was designed and developed.First of all,the purpose and significance of the subject research and the current status of research on visualization technologies in different dimensions as well as the development of marine visualization software at home and abroad are introduced.The visualization algorithms of multiple dimensions are analyzed and summarized.And the 2D scalar,2D vector,and 3D scalar visualization algorithms are selected,which are mainly researched in the subject.Then,the contour generation algorithm based on rectangular grids of 2D scalars,the Line Integral Convolution(LIC)algorithm of 2D vectors,and the 3D texture mapping volume rendering algorithm are mainly elaborated.According to the GPU architecture,the rendering process of these three algorithms are refactored and migrated in parallel,and the high-speed computing performance of the GPU is used to accelerate and optimize the algorithms.At the same time,the visualization algorithms for 2D vectors and 3D scalars are improved.For the LIC texture enhancement algorithm,in the step of using high-pass filtering to enhance the convolutional texture,the texture deformation is avoided by establishing a mapping relationship between the local sampling step and the vector angle instead of selecting a fixed sampling step in the original algorithm,which improves the quality of vector field visualization.For the 3D texture mapping volume rendering algorithm,on the one hand,the color and opacity of the line segment between each two sampling points are calculated by the method of sampling reintegration to eliminate the artifact,which is caused by calculating the value of individual sampling point in the traditional algorithm.On the other hand,the volume data is divided by the octree method to generate data sub-blocks.When traversing the data sub-blocks,the empty sub-blocks containing invalid voxels are skipped and only the valid sub-blocks are operated,which will reduce the amount of calculation and speed up rendering efficiency.Finally,in order to enrich the function for the visualization of global marine environment elements,the 2.5-dimensional spatial surface visualization based on Open Scene Graph(OSG)was proposed.Through the coordinate transformation and color mapping for global ocean reanalysis product data,the visualization of the global distribution of marine environment elements was drawn.Then based on the previous research on multi-dimensional visualization algorithms,the two functional modules of multi-dimensional marine environment element visualization and global marine environment element visualization were integrated.With the addition of human-computer interaction and other functions to improve the experience effect,the multi-dimensional visualization system of marine environment elements was designed and implemented.