【作者】 秦勃；

【导师】 姬光荣；

【作者基本信息】 中国海洋大学 ， 物理海洋学， 2008， 博士

【摘要】 海洋是一巨大复杂、各种因素相互影响,多类过程彼此作用的动态时变系统。海洋科学研究中,为揭示某种海洋现象及其动力机制,确定和找出影响这一现象的海洋要素主要特征结构,即海洋空间数据场中具有特定属性的形状、结构和区域,并加以量化分析、可视化和虚拟仿真是十分关键性的。中国是一个海洋大国,社会和经济发展必然越来越多地依赖海洋。为了迎接海洋21世纪,我国从国家发展战略的高度出发,将“数字海洋”列入“十一五”和国家海洋科学技术发展计划。“数字海洋”是随着“数字地球”战略的提出应运而生的,由海量、多分辨率、多时相、多类型海洋监测监视数据及其分析算法和模型构建而成的海洋信息系统。“数字海洋”采用网络、虚拟现实、科学计算可视化等技术手段,以数字化、可视化、动态显示等方式,模拟重现或预现真实海洋世界的各种状况变化,全面提高我国海洋管理与服务信息化水平,实现开放架构整合和海洋信息共享服务,更好支撑海洋信息应用,无疑具有十分重大的现实意义。本文的工作是863计划“基于网格的海洋环境数据共享与信息服务技术的研究”(No.2006AA09Z139)的重要组成部分,已实现的主要工作和创新如下:1、海洋环境信息可视化网格平台的架构设计为满足海洋环境信息数据可视化在时效性、动态性、精度和规模等方面的处理需求,达到为广大用户提供海洋环境信息数据的远程可视化服务目的,以网格为技术支撑平台实现海洋环境信息可视化网格平台的架构设计。可视化网格平台的体系结构模型分为应用门户层、可视化服务层、网格中间件层和可视化资源层.应用门户层:接收用户的交互信息,生成可视化任务描述文件,向用户显示可视化任务执行结果。可视化服务层:响应可视化应用门户的任务请求,利用可视化网格资源完成具体的可视化任务。其中,可视化任务解析服务负责解析XML格式的可视化任务描述文件,并调用其他可视化服务完成可视化任务。网格中间件层:完成可视化服务的部署、注册、调度、服务发现等网格资源管理功能。可视化资源层:提供网格调度使用的各种可视化资源,包括可视化工具、图形绘制硬件等。可视化网格平台的层次划分清晰、功能定位明确、层间耦合性低,具有很好的可扩展性。2、可扩展的数据格式转换服务针对海洋环境信息的数据格式异构性特点,在分析国内外目前常用的8种海洋环境信息数据格式的基础上,利用XML技术设计并实现了海洋环境信息数据格式转换服务,建立标准统一的格式描述规范和数据格式。数据格式转换服务根据原始元数据和目的元数据的描述特征,通过标准统一的数据格式转换的网格服务,为海洋科学工作者提供方便的数据转换,实现应用数据的无缝自动转换。3、基于物理特征的平面离散流场并行拓扑简化新算法本文提出一种基于物理特征的平面离散流场并行拓扑简化新算法。算法以场的物理特征计算参数为基础,采用图像融合与分割技术从参数图像中抽取满足物理特征的参数进行有选择性的可视化表现。由于物理特征计算分析和流线计算的数据量大、处理耗时,算法实现引入并行处理机制,以提高算法效率、满足远程可视化的需求。实验结果表明,算法在完全保持场的物理特征不便情况下,能有效地对平面离散流场的物理特征进行有选择性地可视化表现,达到流场的拓扑简化与压缩的目的,较好解决了目前流场的拓扑简化与压缩难以并行实现的问题。更多还原

【Abstract】 Ocean is a huge and complex dynamic time-varying system, including various factors interaction with each other. It is very important to detect the special shape, structure and range of the ocean space data, and perform quantitative analysis, visualization and virtual simulation to show certain ocean phenomenon and its dynamics, find the main feature structure of the ocean environment elements that influence the phenomenon.China is an Ocean country, the social and economic development are depended on the ocean more and more. As a strategy of science and technology development, the "Digital Ocean " has been put into the "11th Five-year Plan" and the state ocean science and technology development project, by China government for the 21st Ocean Century. The "Digital Ocean" is proposed with the "Digital Earth". It is an information system composed of large ocean data with multi-resolutions and various types, and also many analysis algorithms and models. The "Digital Ocean" employs many technologies, such as the grid, virtual reality, visualization and etc, to simulate the various states of real-ocean with dynamic, visual and digital display. It is very important to improve the information level of the ocean management and service, to realize the open architecture integration and ocean information sharing service, and to support the application of ocean information better.The work in this paper is an important part of the "863" Project—"Study on Grid based on ocean data sharing and information service" (No. 2006AA09Z139). The following are the main works and innovations.1. The Architecture Design of Grid for Ocean Date Visualization (GODV)In order to meet the real-time and dynamic performance requirement, high precision and large-scale characteristics of ocean information visualization, the architecture of GODV is designed, which can serve the users as a remote ocean data visualization system. The Architecture of GODV is classified into four layers: Grid Portal, Visualization Service, Middleware and Visualization Resources.The application Grid Protal can receive the interactive information from the user, create a file description of visualization task, and display the result of the visualization results to the user. Visualization Service layer can respond to the task application of the Grid Portal, perform the specific visualization task by the Grid Resources. The visualization analysis services can parse the task description file with XML format and call other Visualization Services to finish the visualization task. Middleware performs the Grid Resource management of visualization grid system and its grid service deployment, registration, scheduling, service discovery and etc.. Visualization Resource layer provides various visualization resources for grid scheduling, including the visualization tools, the graphic hardware and etc. Each layer of GODV has the definite function, the relations between layers are less coupled, and produces a more scalable architecture.2. Scalable Data Format Converting services (DFC)As for the heterogeneous data format of the ocean information, DFC is designed and realized based on the analysis of common eight date format types of ocean information and using XML technology. This format converting services use the grid service with uniform data format converting based on the target and the source metadata to meet the ocean research with automatic data format conversion smoothly.3. A New Parallel Algorithm for Physical Feature-Dependent Topological Simplification of Planar Vector FieldThis paper introduces a new parallel algorithm for physical feature-sensitive topological simplification algorithm of planar vector fields. Based on the physical feature computation, the method can detect and selectively visualize the feature meeting the application-specified physical criterion form the parameter image by the image blending and segmentation technology. Since the feature analysis and the computation is complicated and time-consuming, a parallel method is used to improve the algorithm efficiency and meet the remote visualization. The experiment results show that a physical feature preserving topological simplification is achieved perfectly. Meanwhile, the selective visualization of physical feature of planar vector fields is an efficient and remarkable data compression is implemented as well. The method solves the difficulty in parallel computation on topological simplification and compression of flow fields at present.更多还原