海量卫星图像可视化中的多用户协同调度问题研究

发布时间：2018-03-12 13:47

本文选题：协同可视化　切入点：数据调度　出处：《合肥工业大学》2013年硕士论文　论文类型：学位论文

【摘要】：海量卫星图像可视化一直是地理信息系统的热点问题。随着计算机硬件的发展，海量卫星图像可视化在渲染帧率、大数据调度、外存组织等方面产生的传统问题已基本得以解决。在面向多用户的卫星地图可视化方面，Google Earth带来了胖客户端渲染和网络并行传输支持下的多用户解决方案。然而在单机模式下，由于系统处理能力有限，实现面向多用户的卫星图像数据可视化成为了新的挑战。因为图像数据调度算法的优劣会直接关系到最终的场景渲染质量和用户漫游体验，本文研究了目前较为成熟的面向单用户的卫星图像数据调度算法，并在其基础上进行了对比分析，进而针对单机模式下多用户卫星图像数据调度问题提出一种协同数据调度算法，在保证系统数据吞吐量一定的前提下，充分提高系统资源利用率并均衡用户需求，从而实现用户平均满意度的最大化。最后在原型系统上验证了其有效性，并以用户体验方式对其交互性和效率进行统计和分析。本文工作主要包含以下方面： 1）研究面向单用户卫星图像数据调度方案。在传统面向单用户的卫星图像数据调度方案基础上进行改进，根据不同的需求环境设计实现了两种不同的数据调度算法，并通过分析两种可视化框架的优缺点，，为面向多用户的海量卫星图像可视化数据调度问题的解决提供了参考。 2）研究面向多用户卫星图像可视化的协同数据调度方案。针对多用户不用应用背景下请求图像数据的特征对缓冲池结构以及整体渲染框架进行了重构；在此基础之上，提出了一种基于动态优先级原则的协同数据调度算法，不仅能侧重高优先级用户的数据需求也可以保证其他用户的漫游体验；最后，设计协同数据调度算法的匹配实验，并分析实验结果。 3）设计并实现了支持单机模式下多用户同时观察的卫星图像可视化系统。模拟了多用户卫星图像观察的不同应用场景，使用多点触摸交互方式支持单机模式下多个用户的协同操作，在系统吞吐量一定时保证多用户平均满意度的最大化。
[Abstract]:Mass satellite image visualization has always been a hot issue in GIS. With the development of computer hardware, mass satellite image visualization is rendering frame rate, big data scheduling, The traditional problems such as memory organization have been basically solved. In the aspect of multi-user satellite map visualization, Earth brings the multi-user solution with the support of fat client rendering and network parallel transmission. Because of the limited processing ability of the system, the realization of multi-user oriented satellite image data visualization becomes a new challenge, because the quality of the image data scheduling algorithm is directly related to the final scene rendering quality and user roaming experience. In this paper, the mature single user oriented satellite image data scheduling algorithm is studied and compared. Then a cooperative data scheduling algorithm is proposed for multi-user satellite image data scheduling problem in single-machine mode. Under the premise of ensuring the system data throughput, the system resource utilization ratio can be fully improved and the user needs will be balanced. Finally, the effectiveness of the system is verified on the prototype system, and the interaction and efficiency of the system are statistically analyzed in the way of user experience. The work of this paper mainly includes the following aspects:. 1) the single user oriented satellite image data scheduling scheme is studied. Based on the traditional single user oriented satellite image data scheduling scheme, two different data scheduling algorithms are designed and implemented according to different requirement environments. By analyzing the advantages and disadvantages of the two visualization frameworks, the paper provides a reference for solving the problem of visualization data scheduling for multi-user satellite images. 2) the cooperative data scheduling scheme for multi-user satellite image visualization is studied. The buffer pool structure and the overall rendering framework are reconstructed according to the features of the multi-user without the application background. A collaborative data scheduling algorithm based on dynamic priority principle is proposed, which can not only focus on the data requirements of high-priority users but also ensure the roaming experience of other users. Finally, the matching experiment of collaborative data scheduling algorithm is designed. The experimental results are analyzed. 3) Design and implement the satellite image visualization system which supports multi-user simultaneous observation in single computer mode, and simulates the different application scenarios of multi-user satellite image observation. Multi-touch interaction mode is used to support the cooperative operation of multiple users in single-machine mode, and the system throughput is guaranteed to maximize the average satisfaction of multi-users when the system throughput is fixed.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.41;P208

【参考文献】