GPU Based Multidimensional And Dynamic Visualization Of Marine Environment

With the development of 3D visualization technique for the marine-oriented water environment simulation and ocean elements visualized expression, the demand for both the authenticity of marine data visualization and the real-time interactivity is higher and higher in the application field of marine science. Based on Open Scene Graph-the 3D rendering engine and GPU technology, this thesis studies and applies the key technology of marine data multi-dimensional visualization. The results include the following aspects:1.3D visualization Architecture was developed by the use of OSG integrated with MFC framework. This thesis applies building scenes, drawing mechanism, accessor mechanism, callback mechanism and event handling mechanism to realize 3D rendering of the scene. Through the graphical interface development library to develop interactive interface, including floating window、menu bar and status bar, the system controls the expression and feedback of rendering information, to achieve a better human-computer interaction.2. Based on NetCDF data model, this thesis is to realize efficient management to sea water elements. Applying NetCDF, this thesis realizes the efficient and rapid analysis, saving and management of the Multidimensional Oceanic Data, realizes the rapid acquisition and preprocessing of point, line, face and body data and thus to realize the dynamic expression and analysis of sea water element3. Implement efficient rendering of sea water elements based on GPU programming technology. This thesis is to improve scene rendering efficiency and drawing quality after investigating the 3D rendering mechanism and using shading language GLSL. In addition, it realizes the point data analysis, coloring graph of elements and contour tracking on the marine data 3D visualization.4. Realizing the application of volume rendering technology on the 3D marine scene based on the GPGPU technology. This thesis implements the isosurface extraction and Direct Volume Rendering respectively on the basis of Marching CubeS (MC) algorithm and Ray Casting algorithm. These algorithm implementations are optimized in some aspects, such as data sampling, texture mapping, transparency and color composition.5. CUDA programming technology is applied to implement mass data real-time dynamic rendering. In this paper, CUDA parallel computing can effectively control thread activity according to the needs of rendering algorithm. In the process, the GPU completes data calculation instead of CPU.6. Researching and developing a ocean environment multidimensional dynamic visualization software. It integrates OSG and MFC framework and applies the key technology to realize the display of two-dimensional and three-dimensional marine data.