Acquisition of Research Instrumentation for Web-based Visualization of Spatio-Temporal Data

点击翻译按钮获取中文摘要

This project, enhancing database, data delivery, and data visualization in collaboration with the University of Illinois-Chicago, targets the development of technology to address problems in the areas of spatial, temporal, moving objects, and data visualization. Results from the research will be used to enhance TerraFly, an interactive Web-enabled system at http://TerraFly.fiu.edu designed to aid in the visualization of spatial and remote sensed imagery. This research testbed at FIU creates simulated moving object data in real-time and generates overlays of moving objects in real time for display by the TerraFly client. TerraFly, the subject of a cooperative Research and Development Agreement between FIU and USGS, presently takes more than 13 terabytes of aerial photography covering US and makes it available to its users. This mass data, combined with point and polygonal data (made available for TerraFly research by Navigational Technologies) and moving data mined from Internet sources, combined with additional simulated moving object data created by the infrastructure, is expected to allow performance of realistic tests of spatial indexing techniques. Research activities include: Storage and Visualization of Spatial Data Storage and Visualization of Moving Object Data Spatio-Temporal Data IndexingTerraFly includes land, urban, and coastal imagery data collected by satellites, aerial photography and other remote sensing means. The first activity involves enhancements allowing storage of additional metadata, queries, and data processing routines and algorithms, and also to allow on-the-fly incorporation of data stored at remote sites. The second includes handling data update intervals, query languages (linguistic issues), indexing, and uncertainty/imprecision. The third activity performs research toward a new data structure for spatial searching with S-trees (multidimensional search trees optimized for extremely fast updates and queries). The proposed S-Tree is a hybrid data structure that uses a combination of a hash table and a quadtree (a multidimensional binary search tree that divides the space into regions at the points being inserted in the tree).Results of the research will be made available to disaster managers and thus benefit society. Moreover, underrepresented students will be offered hand-on experience for research efforts, thus enabling more potential PhD students in the CS/CE pipeline.