162 The 4th ISPRS Workshop on Dynamic & Multi-dimensional GIS

INTEGRATION OF JAVA AND GIS FOR VISUALIZATION AND ANALYSIS OF MARINE DATA

• NOANNMFS/AFSC, 7600 Sand Point Way NE, Seattle, Wa. 98115, USA and Department of Geosciences, Oregon Stale University, Corvallis, Oregon, 97)31, USA liflimy.c.vDnce@noaa.gov

b JISAO, University ofWasbington, Seattle, Wasbing~n, 98195, USA

KEY WORDS:.marine, three-dimensional, visualization, application

ABSTRACT:

Ideally, scientists should be able to format, explore, anal)lie and visualize data in a simple, powerful and fast appJieatiou that would seamlessly integrate georefereneed data from a variety of data source~ into a powerful intuitive visualization. Geographie information systems (GIS) provide a high level of funetionaliiy for spatial analyses, bill are not yet able to provide the extended funetionality needed to ereate a truly "scientific GIS". Java can be used to program sciemific calculations and analyses, but it isn't inherently spatial. VRML provides the ability to visualize scientific data and to allow the user to interact with the data by rotating. zooming and pllllJ\ing, but you cannot easily query VRML objects. Recent developments in GIS and in Java can be exploited to produce a prototype of this kind of integrated scientific system. In this projeet we used a combinatioo of JavaiJava3D and a commercial GIS package (ArcGIS) to create a prototype for a scientific GIS. We combine the spatial tools exposed through the ArcEnginc API with the analytical capabilities of algorithms wrinen in Java witb the complex visualizllion capabilities of Java3D. Modules from each of these technologies are combiued (0 create iunovative tools to allow users to import data, perfonn spatial and scientific analyses and output the results as visualizations for further examination.

1. INTRODUCTION visualizations, this contrast also snggests there are important bistorical and cultural differences.

1.1 Background Given that ph enomeWl in tbe ocean are inherenUy three­dimensional, it would seem inhJitive that GIS applications for

Ooe of the basie tenets of cartography is that reducing the marine usc should also be three-dimensional. Many of the dimensionaliiy of a phenomenon can be the key to its display, experiments reported in tbe literature are CODducled in Ihree and ultimately to understanding. The spherical world with and four dimensions and analyses are made in multiplethree dimensions is reduced to two dimensions ofa paper map; dimensions. The distribution of tuna ill the Pacific is the four dimensions of oceanographic and aunospberic presented as a function of biotic and abiotic environmeOls phenomena are reduced to three dimensions in a scieutific (Berl(and e\ aI., 2002). Tbe distribution of hydrothermalvisualization. Mauy geog<aphic topics - for example human ven ting on mid,occnn ridge systems is explored by Urabe et aI. settlemem patlerns or the diffusion of ideas or the locations of (1995). Patterns of circulation and tbe variabiliiy of frontal the transcontinel1lal railroads - are inherently four­ locations and dynamics are considered in Hennann et al. dimensional. In most cases, thongh, thcir display and (2002) and Kachel et at (2002), respectively. The marine GIS repres,entation are two- dimensional. The VeGlS while paper literature contains descriptions of the need for truly three­on visualiza.lion (Buckley et aI., 2001) calls for W1ij)ong dimensional GIS (Lockwood and Li, 1995; Yalavanis, 2002; graphics over a variety of dimensions - e.g. 2D points with 3D Fox and Bobbitt. 2000; Meaden, 2000) and theoretical volumes - bUI the research described is mOM: concerned wi th descriptions of what this might entail [Mason et a1., 1994;cognition and usability than the implementation of higher Miller and Kemp, ] 997; Pequet and Wentz., 1994; Kraak and dimensionaliiy. MacEachren, 1994; Yuan, 2001), By eontrast, in oceanography and meteorology, the use of The applied marine GIS literature docs not describe many truly Urrec-dimensional plots and various types of scientific three-dimensional analytical applications. The applicationsvisualizations seems 10 be much more widely a=pted. that do u~e GIS to portray three- and four-dimensional Seientific pnblieations use a varieiy of standard mnlti­ phenomena tend to nse a GIS for data storage and two­dimensional plots, virtual reality is slartiug to be a commonly dimensional calCUlations. Scientific tools such as Matlab or used tool, and the nightly weather report shows animations and custom code for 3D and lime series caleuJatiollS, and a separate multi-dimensional plots. While some of these differences may visualization tool such as Yis5D (Su and Sheng, 1999).simply be due to the types of data collected in the three Fledennans (Goldfinger et aI., 1997), VRML (SAIC, 2003) OF diseiplines, tbe amount of eomputcr processing timc available EYS and VRML (Vance et aI., 2003) are used for the actual to researchers, and the funding available to create advanced visualizations. Part of this is probably because GlS is still

• Corresponding author.

163 The 4th ISPRS Workshop on Dynamic & Multi-dimensional GIS

inherenlly terrestrial and most terrestrial phenomena are 2.5 or 3.5 dimensional. Various authors hove commented on the basic differences between the static terrestriaL environment and the dynamic marine environment (Lockwood and Li, 1995; Wrighl and Goodchild, 1997).

1.2 Motivation and aims

Geographic information systems (G[S) provide a high level of functionality for spatial analyses, but are not yet able 10

provide extended funetionality needed to create a truly "scientifle GIS". Examples of the functionality that is lacking iuclude time series analyses, calculation of the volume of the overlap between two volumes - for example between a school of fish aud a prey lield - or calculation of the intersection of a vector path with a volume - for example the route of a marine mammal through a pool of cold water. Other fuuctions might include the ability to speeify a slice through a three­dimensional lattice of model output data and to make various analyses along thaI sliee. As an example of a 'scientific GIS' I, a test application nsing the project framework was put together to view high-resolution gLobal topographylbathymetry, ocean model output, and standnrd ocean hydrographie data (Figurc I).

Figure I. Bathymetric data displayed as a perspective plot

lava can be used to program scicntific calculations aud analyses, bnt it is not inherently spatial. Datasets can bave a spatial coroponent, but Java !.reats this as it would any type of coordinate system. Topology, or the spatial relationShip between objects, is not stored with datil,' Funetions such as map projections, slope calculations and SpatiaL interseetimlS

_ . By the term "scientific GIS," we mean more than simply adding scientific visnaLization capabilities to an existing GJ§ package. Indeed, we envisage a tool that, fully within a GIS, combines three distiuct capabilities: the traditional spatial tools of a GIS, analylical and computational tools for scientilic analyses in multiple dimensions, and the ability to display the resulls nsing advanced visuaLization techniques.

are not native to the language. However, .lava is easily extensible and functions wmten in other langnage" can be integrated. VRML provides the ability 10 visualize scientific dalll and to allow the user to interact wilh the data by rotating, zooming and panning, but you cannot easily query VRML objects. Ideally one would be able to point at a three dimensional location in a VR1YIL view and return a reference to the ob}ect that IislS the analyli cal methods that acl on the data, and all"""" the user 10 execute these methods. VRML scene navi gab on generally requires that a separate VRML plug-in application be inslalled on the client system. From the user's perspective, plug-ins are poorly integrated iulo the browser environment, and the browser itself may be superfluous. Additionally, user plug-ins may only be available for Ihe more popular client platforms. Recenl developments in Java3D extend the fimctionality of VRML and answer a number of the limLtations mentioned. The Java tools described herein were deslgned using a framework approach that enables Ihem to be integrated into au appLication (OceanGIS), or interfaced with major oft~the-shelf

software products (e.g. ESRI AreEngine). We aimed to develop tbese tools for a broad range of potential users. This project focuses on tools that convcrt non-spatial data into GIS compatible data, expedile the transfer of spatial data to eoastal professionals and emergency managers, and euhanee analyses used for disaster preparedness and response activities. As the tools we develop can be deployed without a filII AreGlS license, we hope to make them widely available to better iotegrate Held activities during disasler responses.

2, METHODS AND RESULTS

2.1 Methods

In this project, we uSe the flexibility of Java 10 iDlegrate GIS functionality with .Tava3D (and Java-wrapped OpenGL) visualization capabilities. Specifically we are using a combinatiou of Java!lava3D and tile reeenlly introduced ArcEugine product to create a prototypc of a scientific GIS. Wc combine thc spatial tools exposed through the ArcEngine Java API with the analylical capabilities of algorithms written in Java with the complex visualization capabilities of lava3D. Modules from these technologies are combined 10 creale innovative lools (0 allow users to import gcoreferenced dara, make spatial seLections, perform spatial and scientific analyses and output the results as vi sualizatious for further examinalion. (Figure 2) Use of the AreIMS .Tava Conneclor will allow these modules to be implemented ill ArclMS site" for web-based analysis. lava allows us to make a variety of scientific calculaLions on the dala and to provide the results both to the GJS component and to a Java3D based visualization component. We are able to lake advantage of a number of Java utilities rueb as Unidata Integrated Data Viewer (IDV)', OceanShare, neBrowse, the

j~.!!!ific..Qf!l~~~_I~()I~it_.{~g.Ir'____"!!.d Ti.meS0-"~"PP1~_ ~.

1 See a description oflDV at hnp:iimy.nnidata.uear.eduieontcntlsoftwarciIDViiudex.btml

, For informationon SGT and the TimeSeries applet see http://www.opic.noaa.goVi]avaJsgt! and

164

The NclMS Java Connector could be used to produce a map coordinates ~base that wonld allow data relrieval from a Distributed Oceanographic Dala Systems (DODS) server" and subsequent plotting with tools designed for illleraction with gridded fields.

Application

Graphical Object (Gob) Creator Bathymetry reader Model output reader CTO database connection

~ Shapefile geometry reader

Gob List Manager Gob

Data Meladata Tools

Gob

Gob ""'"

Gob ._..~~~..~:») Gob ..- ..

/ Rendering/Navigation

Figure 2. Diagram of the components of the OceanGIS system

2.2 Software Components

ArcGlS En[;1ne is an ESRl developer p1'oduct for ereating and deploying ArcGIS solutions. It is a simple API-neutral cross­platfonn development envirO!UTlenl for ArcObj= - tbe C++ component technology framework used t.o build ArcGIS. ArcObjects are the core of the ArcG1S functionality and includ~ tools snch as overlay - union. mterseet; proximi ty ­buffer, point disbnce; surface analysis - aspect, billshade, slope; data conversion - shapefile, coveragc and DEM to gcodaLabase. ArcEugines' object library makes full GIS funetionality available through fine- and coarse-grained components Ihal ean he used in Java and oLhcr cnvironments. Using AreBngine, one can build solutions and deploy them to users without requiring the ArcGIS Desktop applications (AreMap, ArcCatalog) to be present On the same machine. It' supports alltbe standard development environments, ineluding Java, C++, and .NET, and all tbe major operating systems. In addition, one can embed some of the functionality available in the ArcGIS extensions. This product is a developer kit as well as deployment package for ArcO~jecLs technology. Usiug ArcEngine we will integrale GIS functionality into an

hltp:Jlsopae.ucsd.edulproccssinglrefinedJavaTimeSeriesDoc .btml

4 See hltplldQds.l)dbc.no~a.govlfor a description ofDODS.

The 4th ISPRS Workshop on Oynamic & Multi-dimensional GIS

application with the data being available for calculations in non-GIS componen LS. _ The Java3D API is an applicmion programrnmg interface used for writing stand-alone .hree-dimensional graphics applications or Web-based 3D applets. It gives developers high level

•construCtS for creating and manipulating 3D geometry and tools for conslructing the slructures used in rendering rhat geometry. With Java3D API constf\lcts, a'pplication developers can describe very large virtual worlds, which, in rum, are efficienlly rendered by the Java3D API. The Java3D API extension is desigi.'ed BS a high-level platform independenl 3D graphies prograrruning AP~ and is amenable to very high performance implemenr.ations across a range of platfoms. To optimize rendering, Java3D implemenLations are layered to rake advantage of the native, low-level graphics API available on a given system. hi particular, Java3D API implementations are Bvailable that utilize OpenGL, DIfCct3D, and QuickDraw3D. This means that Java3D rendering will be accelerate<! across the same wide range of systems Lhat are snpported by these low-level APIs. The Jnva3D API is aimed at a wide r.mge of 3D-capable hardware and software platfonos, from low cosl PC game cards and software renderers, throngh mid-range workstations, all the way up to very high performance, specialized, 3D image genera~ors.

Support for run-lime loaders was included to allow Java3D to handLe a wide variety of file formats such as interchange fonnats, VRML 1.0, and VRML 2.0. We also explored the use of a second 3D API called the Visualizatiou Toolkit (www.kitware.com). Like Java3D, VTK is a cross-platfonn 3D application programming interface bnill upon, and independent of, the native rendering library (OpenGL, etc). It exposes Java bindings (as well as Tcl and Python). It is written in C++, nses Lbe object oriented modeling approach of Rumbaugh et al. (1991), and includes similar seene-grapb., lightiug models, and graphic primitives as Java3D. What VTK does that Java3D doesn\ (yet) do is boole<Ul operations on 3D volumes (imerseclion, union), volume rendering, filtering, ineluding convolution. J'FT, Gausian, Sobel filters, permutation,' high- and low-pass Butterworth filters, and divergence and gradient calculation. The VTK data mode'! allows for fast lopology traversal, making lhese filters very fit"L, and allows for rapid mesh docimation. VTK also offers powerful 3D probe "widgets" that allow easy interaction with the dar.a, and has methods 10 utilize parallel architeeture through the Mcssage Passiug Interface (MPl).

2.3 Results

For this application (dubbed OceanGlS) we created a number of Java objects that Wl'ap data and metbods for acting on the data. For bathymetry data, this was just lbe ability to e""ggemle the vertical coordinate and to decimate and smooth the resulting mesh. But for more complicated data like hydrographic surveys, the objects allow for the dynamic creation of such rypical oeeanographic parameters as dynamic height, mixed-layer depth, and geostrophic flow. For model output, variables such as salinity can be shown as an isosurface or vector fields sueh as water velocity can be shown BS a plane of 3D vectors that the user ean move to "probe" the data. The user can seed the flow field with lagrangian flO'dts that are time-stepped 10 show partiele paths. In Figure 3, baibymetry from the Smith and Sandwell da1."se1 (1997) are ~elected in the Aleutian Island chain, and the vertical dimension is exaggerated to show tbe Aleutian trench. The image also

165 The 4th ISPRS Workshop on Dynamic & Multi-dimensional GIS

sbows bow a direct link to a DODS data sOW"CC can be included.

Figure 3. Batbymetric data and link to DODS data server

In Figure 4, a em survey just nortb of tbe archipelago is imported in sbapelile format, and mixed·layer deptb is c.alcuJated using a kriging method. Note the data-object icon on Ibe left ponel, exposing 1001 methods to the user . in this case a hydrographic data ~iect was clicked,

Figure 4. Integration of visualization wiil! tool 1.0 ealculate mixed layer deptb and kriged surface

'F;gw:-e' 5" sbows an even higner-resoTiinon'baUiymeme'ilatasef ­(multi-beam), as well as model output of a bydrothennal venl, plume, Dissolved aluminum concemration is shown as a volume, and a plane probe widget slices through the data The user can r..-size the widgel or push it through the volume to

show concentration al that locallon, or to calculate volurne of the plume.

Figure 5. hnage of a bydrothermal plume wilh slieing plane for analysis

One of the challenges in creating a GIS for marine research is linking existing dalabases wilh a GIS. Oceanographers are familiar and comfortable with extracting data from various large archives for immediale analysis. A proeess that involves reformatting these data tor inclusion in a GIS is likely to be seen as cumbersome. Figure 6 shows a direct link between an OPeNDAP server and OceanGIS.

:::.(.: j

-lc~ " =i

:J1r:t~.!.~..~I.~'~:~':~:? ~.-.:~.;:;~:;;;;~}::f:'~~~~~~ _._ ...... ' . ....~~.~.~.\'\~~.,'! ..,~'.'!~xY-o.~.,"'~ ",' .. .,;:;.. : _."~~;"

Figure 6. Using an OPeNDAP server witl) OceanGlS

-DODSJOPeNDAP-~~~;;'~' se~e-ge~;~ie~ced'- d;;;-by"­allowing the user to select dala based on available rneta-data. In lbe marine applicatiolT' shown, hydrographic data are seleeted fim by a simple "rubberbanding" of geographic location and then fine-luning the spatial extent using several

selecti~ tools, and lime aod depth exlepl througltlhe adjacenl

166

panels. This allows lhe user very tine seleclion of large dala sels. Tbe aata transferred back, while not in a standard GIS format (lhey are actually in netCDF fomlal), are georeferenced and can easily be converted. The OceanGIS applicnlion allows 8 uSer to seJoct a datasel using ~ familiar interface and have Ihe data automatically added to the OceanGlS.

I.

3. CONCLUSIONS

Future plans for OceanGIS include the development of marine­SCIence related tools, inclUdlUg lowed-instrument (fence-line) rcndenng, three-dimensional S1atishcal lools such as optimal interpolaliou, and the porting of standard hydrographIc algorithms to the GIS application [UNESCO 1981). The OceanGIS application demonsrr",es the integratlOu of advanced visualization lechmques wilh standard GlS tecbniqnes, and moves the nser from thinking in a twa­dimensional plane to a more )meraC!lve tbree and four­dimensional world. Additionally, the ability to connect to distnbuted data serverS and projeet and visuahu data ou lhe fly opens thIS application to a new gronp of users, including emergency managers and scielHific modeller.; who must respond 10 disaSler managemelll seenarios rapidly. '>'-'hile llone of tbe various technologies we used is complete in and of ilsclf, Ihe linking of GIS, Java, .Iava3D and VTK provides a powerful mechanism to Create the beginnings of a "scientific GIS". Once the technologies have been linked, the form of lhe tillaL tools deployed will be dependent upou the users needs and the relative costs of the GIS portions of the technologies. Because we are creating seientijic 1001s rather than a <:omrnercial prodnct, license c<:>sts may have to playa large role in our choice of which GIS tools to use.

4. REFERENCES

BcrLIand, A., Josse, E., Bach, P.. Gros, P., and Dagorn. L. 2002. Hydrological and Trophic Characlcristics of Tuna Habitat: Consequences On Tuna Disnibuliou and Longline Catchability Canadian Journal of Fisheries and AquGlic Sciences, 59(6l,pp.1OO2-1013.

Buckley, A. H. Gahegan and K Clarke. 200 I. Geographic V,sualization. UCGlS white paper http://www.ucgis.otglprioritieslresearchlresearch_white/2000w hitepapersinde". Accessed 12 July 2005.

Fox, C. and Bobbitt A. 2000. The National Oceanic and Atmospheric Admjllls1ration's Vents Program GIS: Integration, Analysis and Dis(ribution of Mnltidiseiplinary Oceanographic Data. In Wright, D. and Bartlctt D. Morine and Caa<tal Geographical Information Systems. London: Taylor and Francis.

Goldfinger, C. L. McNeill and C. Hnmmon. 1997. Casc Study of G1S Dara lI1tegr-atiou and Visualization in Manne Tectonics: The Cascadia Snbdu<.iion Zone. Martne Geodesy, 20, pp. 267-289.

Hennann, A. J., Slabeno, P. J., Haidvogel, D. B., and Musgrave, D. L. 2002. A Regional Tidal/Subtidal Circulation Model of the Southeastern Bering Sea: Development, Sensitivity Analyses and Hindeasling. Deep-

The 4th ISPRS Workshop on Dynamic & Multi-dimensional GIS

Sea Research Part I -Toptcol Srudies in Oceanography, 49(26), pp. 5945-5967.

Kaehel, N. B. Hunt G. L. Salo S. A. Schumacher J. D. Slabeno P. J. and Whitledge T. E. 2002. Characteristics and

, variabililY of the inner front of the southeaslern Bering Sea. Deep-Sea Research Part JI, 49, pp. 5889-5909.

Kraak, M. 1. & A. M. MacEachren 1994. Visualization of spatial data's leroppral component., 111 Waugh, T. C. & R. G. Healey Advances' in GIS Research. 391-409.

Lockwood, M. and Li R. 1995. Marine Geographic infmmallon Systems What Sets Them Apart? Morine Geodesy, 18:157-159.

Mason. D. M. O'Conaill and S. Bell. 1994. Handhllg four­dimensional ge<>-references data in envIronmental GlS. lntemational Journal if Geographical lnfarmation Sysrerns, &( 2), pp. 19/-215.

Meaden, G. 2000. Applicatious of GIS to Fisheries Management, iu Wrigbl, D. and Bartlett D. Marine and Coastal Geographical lllfonnalion Systems. London: Taylor and FranCls.

Miller, E. Kemp, Z. 1997. Towards a 4D GIS: Four­dimensional interpolation utilizing krigmg. In Innovations in GIS 4. London: Taylor end Francis; 181-197.

Pequct., D and Wentz E. 1994. An Approaeh for Timed-Based Spatial Analysis of Spalia-Temporal Dala. hI Advances in GIS Reserl/'ch, Proceedings 1:489-504.

Rumbaugh 1.. Blaha M., Fremerali W., Eddy F., and. Lorensen W. 1991. OojeC/-Oriellled Modelling and Design. Englewood Cliffs, New Jersey: Prentice-Hall.

SAlC. 2003. Techniqnes for spatial analysis and visualizauon of benthic mapping data. Report for NOAA/CSC.

Smith W.H.F. and D.T. SandweH, j 997, Global Sea Floor Topography from Satellite Altimctry and Ship Depth Souudmgs, Science. 277 (5334). pp. 1956-1962.

Su Y. and Sheng Y. 1999. Visualizing upwelling a( Mont.ercy Bay in an integra ted envirorunenl. of GIS and scientific visualization. Marine-Geodesy, 22(2), pp. 93·104.

UNESCO 1981. The Practical Salinity Scale 1978 and lhe International Equation of Stale of SewaLer 1980. Tenth Report of the Jon: I Panel Oll Oceanographic Tables and Standards (lPOTS, Sidney, B.C., Canada, 1-5 September, \980. UNESCO Technical Papers in Marine Science No 36, 25pp.

Urabe, T., Baker, E. T., lshibasbi, .I., Feely, R. A, Mammo, K., Massot.h, G. J., Maruyama, A., Shitashima, K., Okamura, K., LnproD, J. E., Sonoda, A., Yamazaki, T., Aoki, M., Gendron, J., Greene, R., Kaiho, Y, Kisimolo, K., Lebon. G., Matsumoto, T., Nakamura, K., Nishizawa, 1\., Okano, 0., Paradis, G., Roe, K., Shibata, T., Tennant, D., Vance, T., Walker, S L., Yabnki, 1.. and Ylow, N. 1995. The Effect of Magmatic AClivi[)' On Hydrothermal Vcntmg Along lhe Superfast-Spreading East Pacific Rise. Science, 269(5227), pp. 1092-1095.

167

Valavanis, Y. 2002. Geogrophic Infonnarion Sy,rrems in Oceanography and Fisheries. London: Taylor and Francis.

Vance, T. C. Merati N. Moore C.•nd Alexander C. 2003. VRML-based visualization of GIS data for a marine sanctuary. In Proceedings afrhe 19th Conference on IJPS, AMS 2003.

Wright, D. and M. Goodchild. 1997. Data from the deep: Implications for the GIS community. Inlemational .Journal of Geographic jnformolion Science, 1I (6), pp.23-51S.

Yuan, M. 2001. Represcnung Complex Geographic Phenomena wllh both ObJect- and Field-like Properties. Car'/ography and Geographic 1nfomiGfion Sci(mce, 28(2), pp. 83-96.

PMEL scientilic visualization website: hlrp:I!Wv,-v.,·.pmel.noaa.gov1vrIlO""auGlS Accessed 12 July 2005.

VTK websile: hitp://www.kitwar~ COJI1,/VKT Accessed 12 July 2005.

Java3D website: htll'y iJava.sun.com!prodnctsl.iava-me(.\in/3D Accessed 12 July 2005.

5. ACKNOWLEDGEMENTS

Many t1,.nk, 10 ,he two anonymous reviewers for helpful comments. funding for this projcct was provided by the High Performance Computing and Communications (HPCC) projOCl oflbc NOAA Office of the Chief Information Offi""r. For more details please see http:hrn.I,"\~...eb.nwTl.noan.gov!hpcc/nwg/. Additional support for this research was provided by an NSF IGERT graduale fetlowship (NSF award 0333257) in the Ecosystem lnfonnaties IGERT program al Oregon State University. 'TI,is research is contribution FOCI-0563 to NOAA's Fisheries-Oceanography Coordinated Investigations. This publication was supported by the Joint Institute for the Study of the Atmosphere and Oceau (JISAOj under NOAA Cooperative Agreement #NAl7RJ1232, Contribution 1144. PMEL eontribution 2830. The views expressed herein are [hose of tbe anlhor(s) and do not necessarily reflecl the views of NOAA or any of its sub­agencies. Mention of software products does noi iIiJply cn(.\orsement of these products.

·"...---.:.;-=1 International Society for Photogrammetry and Remote Sensing i~/-~~~i~J:l~ ,Societe Internationale de Photogrammetrie et de Teledetection . .1:"1/

.~.:.._._.,""'"

Internationale Gesellschaft fUr Photogrammetrie und Fernerkundung ~-~

'~(b""'~U- ,t .... '_Pl'1

I.

C THE .INTERNATIONAL ARCHIVES OF THE PHOTOGRAMMETRY, REMOTE SENSING AND SPATIAL INFORMATION SCIENCES

ARCHIVES INTERNATIONALES DE PHOTOGRAMMETRIE, DE TELEDETECTION ET DE SCIENCES DE L'INFORMATION SPATIALE

INTERNATIONALES ARCHIV FOR PHOTOGRAMMETRIE, FERNERKUNDUNG UNO RAUMBEZOGENE INFORMIITIONSWISSENSCHAFTEN

VOLUME PART • VOLUME TOMEXXXVI 2/W29

BAND TElL

The 4th Workshop on Dynamic & Multi-dimensional GIS

Pontypridd, Wales, UK

5 - 8 September 2005

Editor Christopher Gold

Organiser ISPAS Inter Commission Working Group III IV

"Dynamic and Multi-dimensional Systems and Applications"

Sponsors University of Glamorgan

Ordnance Survey National Geomatics Center of China

EU Marie Curie Actions