Research On Dynamic Visualization And Real-time Rendering Technology For Massive Marine Data

China is a major maritime country. In the context of “sea power” strategic decision, Itis important in the construction of the digital ocean to provide comprehensive andaccurate marine information and product and to integration, display, refining andanalysis the effective information. The ocean data is the basis for building a digitalocean. It is the most intuitive and effective means for user to express the marineinformation and data through the scientific visualization techniques. Currently, marinedata in multiple formats and the explosive growth of the amount of information thatrequire giga byte size. How to take full advantage of the performance of the existinghardware and propose algorithms to meet the large-scale scene rendering speedbecomes dynamic visualization for massive ocean data research focus. This canreduce the impact of internal and external data stored in the roaming process a largenumber of exchange frame rate.This paper designs the management of scene and rendering engineering in thei4Ocean for the needs of the marine areas. The main research contents of this paper isto develop an internal and external memory scheduling architecture that neglect theupper application semantics, and on this basis to implement the visualization formassive marine data. The main contents and results of the research in this dissertationare summarized as follows:(1) The analysis and design of marine visual dataWe analyze the types and formats of marine data, and design the data structure basedon the multi-source heterogeneous. This can maximize the support of existing marinedata and solve the multi-format compatibility problems between the marine data.(2) The implementation for scene architecture and management for draw resources fori4OceanThis paper has made a contribution on the scene management of i4Ocean, including the management for scene object and draw resources. For data management, wesupport the unified management of different types of scene data and provide standardmanipulation interface. We also manage scene nodes using the type pool in order tofacilitate the process of debugging. It improves the speed of file access to use binarystream for data storage. For the management of draw resources, we encapsulateOpenGL state and draw all scenes through the programmable pipeline.(3) The design and implementation of marine data visualization systemRelying on the marine data structure that we design and the MVC framework, webuild a visualization system for multi-dimensional marine data. The system is able toachieve the visualization for contour, color section, isosurface and vector arrow.(4) The research and implementation of an internal and external memory schedulingarchitecture based on block strategyWe develop an internal and external memory scheduling architecture that neglect theupper application semantics for the rendering of massive data. To reduce the cost ofIO operations, we research algorithm on incremental saved after scenes updated. Andwe also need to control the number of data fragmentation in the hard disk file.This paper implements the visualization for massive ocean data based oni4Ocean system. The out-of-core system is able to support the multidimensionalvirtual interactive experience between marine environment and landscape, and meetthe needs of visualization for massive marine data. The research can provide thevisual simulation and decision support for the marine complex environmentoperations, and timely response to external information and changes in theenvironment. It can be used to predict environmental changes of the land and sea aswell as multi-dimensional visual experience.