SIGGRAPH 2004 Course #30 Notes
Visualizing Geospatial Data
Topic #1: Overview of Integrating Geospatial Data with Visualization Methods
Theresa-Marie Rhyne
tmrhyne@ncsu.edu
Learning Technology Service
North Carolina State University
and
Director of the ACM SIGGRAPH Carto Project
rhyne@siggraph.org
Introduction:
This discussion provides an overview of six aspects of integrating geospatial data with visualization methods: (1) the Evolution of GIS and Visualization (Vis); (2) Integrating GIS and Vis (SciVis & InfoVis) Tools; (3) Virtual GIS and World Wide Web (Web) Developments; (4) Time Series Animations; (5) Handheld & Wireless Computing Considerations and (6) Commonplace Visualizations.
Figure #1:Typical Scientific Visualization of air pollution concentrations. Computational model data is filtered and mapped into geometric primitives. A geographic map is used to provide context for the region under study (the Northeastern part of the United States of America). This visualization was developed from 1990 - 1992 using rudimentary methods of geographic information system (GIS) and visualization (Vis) integration. These course notes will later describe the levels of GIS and Vis integration. Image created by Theresa-Marie Rhyne while working for Unisys Corporation at the United States Environmental Protection Agency (US EPA), using visualization software developed by Lee Westover while at Numerical Design, Ltd. (on a contract with Unisys Corporation in support of the US EPA).
The Evolution of GIS and Visualization (Vis):
A Geographic Information System (GIS) is frequently defined as the combination of a database management system, a set of operations for exploring data, and a graphic display system that are tied to geospatial analysis of data. GIS environments are also cartographic tools that facilitate building maps in real time and examining the impacts of changes to the map interactively.
Scientific visualization (SciVis) converts numerical or symbolic data and information into geometric computer generated images. It is a methodology for interpreting image data entered into a computer as well as data generated from computational models. Generally, SciVis is based on the application of techniques from the convergent fields of: computer graphics; image processing; computer vision; computer-aided design, signal processing, and user interface design. SciVis research and development has focused on issues pertaining to three-dimensional computer graphics rendering, time series animation, and interactive (in real time) displays via computers.
In the late 1980’s and early 1990’s, both of these disciplines evolved in parallel to each other. Efforts to develop geospatial data standards rarely included how to visualize the data. Computer graphics rendering libraries and standards evolved independently of geospatial data models. This resulted in inefficiencies associated with geovisualization. These include difficulties with registration of geospatial data within SciVis software, cumbersome productions of animation sequences within GIS environments, and perhaps more importantly, the lack of connections between the database and the visualization environment that supports display of geospatial data reliably.
Reference: Theresa Marie Rhyne, William Ivey, Loey Knapp, Peter Kochevar, AND Tom Mace, "Visualization and Geographic Information System integration: what are the needs and the requirements, if any??", in Proceedings of the IEEE Computer Society Visualization 94 Conference, October 17 - 21, 1994, Washington, DC, IEEE Computer Society Press, Los Alamitos, California, 1994, pp. 400-403.
Figure #2: Early attempt at integrating GIS (ArcInfo) environments with SciVis (AVS) tools. We depict here a simple AVS network for importing ARC TIN files. Image produced in 1996 and created by Theresa-Marie Rhyne and Thomas Fowler while working for Lockheed Martin Technical Services at the United States Environmental Protection Agency (US EPA)'s Scientific Visualization Center.
As Information Visualization (InfoVis) evolved and matured in the mid to late 1990’s, geographers and cartographers began to actively participate in this new arena of visualization. InfoVis tends to focus on examining visual metaphors of non-inherently spatial data such as text, hierarchies, and statistical elements. Cartographic methods were and continue to be applied to depict non-inherently spatial data.
Figure #3a:Information Visualization of topics or themes within a set of documents depicted as a relief map of natural terrain. This visualization concept, entitled ThemeView, was developed at Pacific Northwest National Laboratory as part of the SPIRE - Spatial Paradigm for Information Retrieval and Exploration - tool, (http://www.pnl.gov/infoviz/spire/spire.html). Image shown courtesy of Pacific Northwest National Laboratory which is managed and operated by the Battelle Memorial Institute on behalf of the United States Department of Energy.
Figure #3b:Information visualization of Multicast Backbone Internet traffic using a 3D globe to represent the data. This visualization is part of the CAIDA toolset for network drawing that was written by Tamara Munzner and Eric Hoffman. See: (http://www-graphics.stanford.edu/papers/mbone/). Image shown courtesy of Tamara Munzner, currently at the Department of Computer Science, University of British Columbia.
At the same time, focus groups were formed to examine specific issues relating to visualizing geospatial data. The International Cartographic Association (ICA) formed the ICA Commission on Visualization in 1995 to address core research issues associated with extending cartographic methods into the visualization arenas. In 1999, the commission was renamed to the ICA Commission on Visualization and Virtual Environments, (http://www.geovista.psu.edu/sites/icavis/). In 1996, the Association for Computer Machinery’s Special Interest Group on Graphics (ACM SIGGRAPH)’s Carto Project was formed to examine how viewpoints and techniques from the computer graphics community can be effectively applied to cartographic and spatial data sets, (http://www.siggraph.org/~rhyne/carto/). In 1998, the GeoVRML Working Group of the Web3D Consortium was formed to develop tools and recommended practices for the representation of geographical data using the Virtual Reality Modeling Language, (http://www.geovrml.org/). VRML and GeoVRML will be highlighted in later sections of these course notes. These three groups have collaborated together over the last six or more years to examine methods for visualizing geo-referenced and geospatial data.
In the 2000’s, geovisualization has emerged as its own unique subfield with its own research challenges and agenda. Four major aspects of geovisualization include: (a) representation of geospatial information; (b) integration of computational and visual geographic methods; (c) creation of effective interface designs for geovisualization tools; and (d) the study of the usability of geovisualization environments. In January 2001, the International Cartographic Association’s Commission on Visualization & Virtual Environments published a geovisualization research agenda in the journal of Cartography and Information Science. These efforts involved collaboration with the ACM SIGGRAPH Carto Project.
Reference: Research Challenges in Geovisualization, Cartography and Geographic Information Science, vol. 28, January 2001, (A. M. MacEachren and M.-J. Kraak, editors), see Guest editors introduction, pp. 3 – 12.
Integrating GIS and Vis (SciVis & InfoVis) Tools:
In the late 1990’s and continuing into the 2000’s, strides were made to integrate GIS and Vis tools. GIS developers have explored how to incorporate three-dimensional and time series animation capabilities into their software. For example, in the late 1990’s, ESRI introduced ArcView 3D Analyst to allow for visualizing surface data and three dimensional modeling. Today, ArcGIS 3D Analyst is integrated into the ArcGIS suite of tools, (http://www.esri.com/software/arcgis/arcgisxtensions/3danalyst/index.html). Meanwhile, SciVis and InfoVis programmers built data readers that support geospatial data formats. As an example, Advanced Visual Systems (AVS), in the late 1990’s, developed an AVS-Arc data reader that allowed for direct operational import of ESRI’s Arc-Info data into the AVS visualization environment. Today, GIS import and database interaction are incorporated into AVS/Express Professional Edition, (http://www.avs.com/software/soft_t/specs.html#XPV).
Figure #4: Image of a virtual community with air pollution (Ozone) tracking data, built with ArcView 3D Analyst in 1999. This work was done as part of a "Human Exposure in Urban Environments" project for the United States Environmental Protection Agency (US EPA) - Alan Huber, principal investigator. Image created by Theresa-Marie Rhyne while working for Lockheed Martin Technical Services at the US EPA Scientific Visualization Center.Richard Greene and Dick Dulaney (GIS - Arc/Info & ArcView experts) of the Lockheed Martin Remote Sensing Team, Bettina Brinkley (EPA-ETSD Intern with a Geography background), and Robert Lin of the Lockheed Martin - US EPA Sci. Vis. Center team provided technical input for the execution of this visualization.
A GeoVRML example of Figure #4 is available at: (http://www.siggraph.org/~rhyne/vis-tut/gisvis-tut/track-feet.wrl). A VRML plug-in to your Web Browser is required for this viewing. Parallel Graphics provides a VRML plug-in called the Cortona VRML Client. To download this free VRML plug-in, go to (http://www.parallelgraphics.com/products/cortona/).
Figure #5: Three-dimensional texture image of a USGS 100K series topographic map shown in the AVS-ARC system with the two dimensional map shown in another background window. Research conducted by Theresa-Marie Rhyne and Thomas Fowler while working for Lockheed Martin Technical Services at the US EPA Scientific Visualization Center in 1995.
In examining such efforts, four levels of GIS and Vis integration can be defined:
- Rudimentary: minimal data sharing between the GIS and Vis systems
- Operational: consistency of geospatial data
- Functional: transparent communication between the GIS and Vis systems
- Merged: one comprehensive toolkit environment
AVS/Express Professional Edition provides rudimentary and operational GIS and Vis integration. Today’s version of ArcGIS 3D Analyst provides functional integration between the GIS and Vis components of the ArcGIS environment. Many InfoVis tools, while often incorporating only two dimensional (2) visual displays, have succeeded at approaching merged GIS and Vis integration. In the computer games arena, real-time terrain engines are designed to address rudimentary integration.
Virtual GIS and World Wide Web (Web) Developments:
The introduction of the Virtual Reality Modeling Language (VRML) in 1984 provided for interactive three dimensional (3D) representation of content on the World Wide Web, (http://www.web3d.org/x3d/spec/vrml/VRML1.0/index.html). After three or more years of community effort, VRML 97 was approved by the International Organization for Standardization (ISO) and the IEC (the International Electrotechnical Commission) as an open file format for describing three-dimensional (3D) objects and worlds via the Internet. Information on VRML 97 can be found at: (http://www.web3d.org/x3d/spec/vrml/vrml97/index.htm). During the same time frame of the development of this standard, the VRML Consortium was formed to foster the continued development VRML. The VRML Consortium was charted in early 1997 and changed its name to the Web3D Consortium in December 1998 to address the standardization of multiple technologies associated with 3D on the Internet, (www.web3d.org).
Figure #6: Early example of a VRML 1.0 file created with an (AVS to VRML 1.0) Module. The AVS module was developed by John Evans and Richard Signell at the U.S. Geological Survey. At the U.S. EPA Scientific Visualization Center, we then combined the (AVS to VRML 1.0) module with our AVS-ARC networks. This allowed us to pipe geospatial data directly from a GIS environment into a visualization system and output to the VRML 1.0 format. Research conducted by Theresa-Marie Rhyne and Thomas Fowler while working for Lockheed Martin Technical Services at the US EPA Scientific Visualization Center in mid-1996.
An early example of the application of VRML to Geographic Data Sets can be found at: (http://www.siggraph.org/~rhyne/carto/mariposaw2.wrl). This data for this example came from the US Geological Survey as a Digital Elevation Model. The geographic location is Mariposa, California.
As noted in the previous section of this writeup, the GeoVRML Working Group of the Web3D Consortium was chartered in February 1998 to facilitate the viewing of geo-referenced data, like maps and 3D terrain models, over the Web via VRML plugins for Web browsers. By 2000, this resulted in the production of a specification and open source code, entitled GeoVRML, for representing and visualizing geographic data using VRML 97, (http://www.geovrml.org/1.0/).
Reference: M. Reddy, L. Iverson, and Y. G. Leclerc (2000). "Under the Hood of GeoVRML 1.0". In Proceedings of The Fifth Web3D/VRML Symposium. Monterey, California. February 21-24, 2000. (http://www.ai.sri.com/~reddy/pubs/pdf/vrml2000.pdf).
Figure #7:Snapshot of a GeoVRML visualization. The original geovisualization was created with ArcView 3D Analyst in 1998, exported to VRML97 format, and enhanced with GeoVRML techniques in 1999 and 2000. This visualization was developed for the US EPA's Human Exposure in Urban Environments Project - Alan Huber, principal investigator. Image created by Theresa-Marie Rhyne while working for Lockheed Martin Technical Services at the US EPA Scientific Visualization Center.
In 2001, the next generation VRML standard was introduced and entitled Extensible 3D (X3D), (http://www.web3d.org/x3d/spec/ISO-IEC-19775/index.html). X3D was designed to improve upon VRML with new features, advanced application programmer interfaces, additional data encoding formats, stricter conformance, and a component architecture that facilitates a modular approach to supporting the X3D standard. GeoVRML functions were incorporated into the Geospatial profile of X3D with the intent of providing accurate placement and rendering of objects in a 3D Geospatial context. A tutorial on the Geospatial profile of X3D was presented at the Web3D 2004 Symposium (April 2004) by Mike McCann, (http://www.web3d.org/s2004/tutorials.html#Geo).
GeoVRML functions have also been included in OpenGIS Consortium (OGC) discussions regarding Web Services for geospatial data. The OGC is a non-profit member-driven organization aimed at fostering the development of geoprocessing interoperability computing standards, (http://www.opengis.org). Research is currently underway in the OGC regarding missing interoperability functions of three dimensional geovisualization components.
Reference: Angela Altmaier and Thomas H. Kolbe, “Applications and Solutions for Interoperable 3d Geo-Visualization”, Proceedings of the Photogrammetric Week 2003 in Stuttgart, Wichmann Verlag (D. Frissch, editor), (http://www.ikg.uni-bonn.de/kolbe/publications/Altmaier_und_Kolbe_PhoWo2003.pdf).
Time Series Animations:
Traditional GIS environments frequently have not incorporated tools for building time series animations. As a result, other tools like Quicktime and Flash are often used to create the animations from the GIS imagery. For example, using ArcGIS 3D Analyst, each frame of a 24 time step sequence can be stored as a JPEG image. Using the Layout editor, metadata elements like notations of each individual time step, a map legend and titles for the animation sequence can be added to each JPEG image. All 24 frames are then be assembled into a Quicktime or Flash movie for viewing on the Web across the Internet. More complex “rich media” presentations that include audio narrations can be developed using streaming media technologies like the Synchronized Multimedia Integration Language (SMIL), (http://www.w3.org/AudioVideo/).
Reference: Ian Johnson and Andres Wilson, “The TimeMap Project: Developing Time-Based GIS Display for Cultural Data”, Journal of GIS in Archaeology, Vol. 1, ESRI Inc., Redlands, California, 2002, (http://www.esri.com/library/journals/archaeology/volume_1/time_based_display.pdf), Or: (http://www.timemap.net/documents/publications/2000_j_gis_arch_esri/timemap_article/index.html) .
Reference: Theresa- Marie Rhyne, Web Horizons for Geographic Visualization, Geoinformatics, December 2000, pp. 35-37.
Figure #8:Image from a 24 time step QuickTime movie animation sequence of a Mobile Emissions computational model for the Wake County, North Carolina (USA) area. Each individual frame was created in ArcView 3D Analyst and assembled into a QuickTime movie in 1999 and 2000. This work was done as part of the Mobile Emissions Characterization Visualization Project for the United States Environmental Protection Agency (US EPA) - Sue Kimbrough, principal investigator. Image created by Theresa-Marie Rhyne while working for Lockheed Martin Technical Services at the US EPA Scientific Visualization Center.
A Quicktime Movie of Figure #8 is available at: (http://www.siggraph.org/~rhyne/vis-tut/3dcolayout.qt). A Quicktime Movie plug-in for your Web Browser is required for this viewing. Apple provides their free Quicktime Movie plug-in at (http://www.apple.com/quicktime/products/qt/).
Handheld & Wireless Computing Considerations:
Global positioning systems (GPS) are currently available on mobile, handheld platforms such as personal digital assistants (PDAs) and Cellular phones. Other cartography and mapping applications have also been ported to these small screen devices. There has also been success in porting VRML and GeoVRML to handheld devices. Pocket Cortona, from Parallel Graphics, allows for viewing VRML scenes on wireless devices such as the PocketPC, (http://www.parallelgraphics.com/products/cortonace/). As a result, GeoVRML applications can also be ported to wireless handheld devices.
Figure #9:Original visualization by Michael Holmes, Dr. John Fels and James Tomlinson of North Carolina (NC) State University’s College of Design. Enhanced GeoVRML Visualization by Theresa-Marie Rhyne, NC State University – Learning Technology Service (circa 2002/2003).
Commonplace Visualizations:
The recent dominance of computer games in the entertainment arena has resulted in significant impacts on the evolution of computer graphics hardware, software, image rendering, and virtual reality. This has also increased the use of commodity graphics boards in high-end scientific applications like visualization. Yielding the realization that scientifically reliable visualizations will likely be eventually performed on computer game consoles as well as wireless PDAs and cell phones. This is also true for geovisualization and the resulting visualization of geospatial data.
Reference: Theresa-Marie Rhyne, “Computer Games and Scientific Visualization”, Communications of the Association for Computing Machinery (CACM), Vol. 45, No. 7, July 2002, pp. 40 – 44.
Figure #10:Visualization of a virtual community shown on a mobile cell phone. Image shown courtesy of Lars Bishop and David Holmes of Numerical Design Limited, (http://www.ndl.com).
Computer games developers have long used terrain databases, publicly available from the United States Geological Survey (http://library.usgs.gov/), as starting points for terrain modeling of scenes. Urban Planners have realized the advantages of developing community involvement activities that include interactive elements derived from popular computer games like SimCity (http://simcity.ea.com/) and The Sims (http://thesims.ea.com/). As a result, we are approaching a juncture where the software for creating 3D geovisualizations and Web access to geospatial repositories will be widely available online to the general public. This will result in commonplace interactive visualizations created by general users of desktop computers. We can see these activities starting to take place in arenas such as traffic planning and engineering.
Figure #11:Typcial traffic visualization example. Image shown courtesy of North Carolina Department of Transportation, (image created by Chris Parker). For more information on this project, see: (http://www.ncdot.org/projects/Superstreet/).
It will be the task of geovisiualization professionals to champion the integrity, reliability and usability of online visualization methods for geospatial data. This involves inter and intra disciplinary collaborations with colleagues in computer graphics, cartography, geographic information systems, telecommunications infrastructure, mobile computing, distance education, computer games and many other application disciplines.
Acknowledgements:
We thank Alan Huber and Sue Kimbrough at the United States Environmental Protection Agency (U.S. EPA) for providing projects that tested the early limits of geographic visualization as well as Thomas Fowler for several years of collaborative work at the U.S. EPA Scientific Visualization Center. James Tomlinson, Michael Holmes, and John Fels provided unique geovisualization and geovrml applications from North Carolina State University (NCSU)'s College of Design. Pak Wong, Kristin Cook, and David R. Cook of Battelle Memorial Institute at Pacific Northwest National Laboratory provided input regarding ThemeView and other information visualization techniques. Tamara Munzner, currently at at the Department of Computer Science, University of British Columbia, also provide insight on information visualization and network drawing approaches that apply cartographic methods. Lars Bishop and David Holmes of Numerical Design Limited contributed approaches to the use of geovisualization in computer games applications and on handheld devices. Chris Parker and James H. Dunlop from the North Carolina Department of Transportation provided input on the development of common place visualization for highway and local traffic design. Additional thanks to Alan MacEachren, Pennsylvania State University, for collaborating with me on this "Visiualizing Geospatial Data" course at SIGGRAPH 2004 and to the SIGGRAPH 2004 Courses Committee (Jacquelyn Martino, Chair) for selecting our course proposal.
Summary of References:
Theresa Marie Rhyne, William Ivey, Loey Knapp, Peter Kochevar, and Tom Mace, "Visualization and Geographic Information System integration: what are the needs and the requirements, if any??", in Proceedings of the IEEE Computer Society Visualization 94 Conference, October 17 - 21, 1994, Washington, DC, IEEE Computer Society Press, Los Alamitos, California, 1994, pp. 400-403
Research Challenges in Geovisualization, Cartography and Geographic Information Science, vol. 28, January 2001, (A. M. MacEachren and M.-J. Kraak, editors), see Guest editors introduction, pp. 3 – 12.
M. Reddy, L. Iverson, and Y. G. Leclerc (2000). "Under the Hood of GeoVRML 1.0". In Proceedings of The Fifth Web3D/VRML Symposium. Monterey, California. February 21-24, 2000. (http://www.ai.sri.com/~reddy/pubs/pdf/vrml2000.pdf).
Angela Altmaier and Thomas H. Kolbe, “Applications and Solutions for Interoperable 3d Geo-Visualization”, Proceedings of the Photogrammetric Week 2003 in Stuttgart, Wichmann Verlag (D. Frissch, editor), (http://www.ikg.uni-bonn.de/kolbe/publications/Altmaier_und_Kolbe_PhoWo2003.pdf).
Ian Johnson and Andres Wilson, “The TimeMap Project: Developing Time-Based GIS Display for Cultural Data”, Journal of GIS in Archaeology, Vol. 1, ESRI Inc., Redlands, California, 2002, (http://www.esri.com/library/journals/archaeology/volume_1/time_based_display.pdf), Or (http://www.timemap.net/documents/publications/2000_j_gis_arch_esri/timemap_article/index.html) .
Theresa- Marie Rhyne, Web Horizons for Geographic Visualization, Geoinformatics, December 2000, pp. 35-37.
Theresa-Marie Rhyne, “Computer Games and Scientific Visualization”, Communications of the Association for Computing Machinery (CACM), Vol. 45, No. 7, July 2002, pp. 40 – 44.
----------------------------------------------------------------------------------