Multi-frame and multi-dimensional historical digital cities: the Como example

Abstract

In this article, we present the realisation of a multi-frame and multi-dimensional WebGIS that allows users to simultaneously analyse a specific portion of the Earth taking into account the historical information, too. Two graphical panels have been realised: one for the usual 2D view and one for a more realistic 3D view. Both panels display historical maps of the city, the current orthophoto and the digital topographical map. The 3D frame is based on NASA World Wind, an open source virtual globe from where 3D buildings are shown extruding the 2D shapes using their mean height. Thanks to a specifically designed graphical user interface, it is also possible to dynamically thematise the buildings on the globe according to different criteria (e.g. the construction time span) so that only the geometries fulfilling the request are turned on. Within the proposed application, a synchronisation between the two panels has been implemented, in order to maintain a constant alignment of the two viewers. The application is also open to the time dimension. In fact, assigning to each geometry two dates (e.g. ‘year of construction’ and ‘year of demolition’), it is possible to dynamically view how buildings have changed over time, both in their shape and height. Future developments of this work will concern the possibility of implementing a city model with a higher level of detail.

Keywords:

• web application
• multi-frame
• multi-dimensional
• time
• historical maps
• 3D

1. Introduction

The purpose of this article is to contribute to the advanced Web-mapping, allowing the synchronisation of 2D visualisation of features varying in time with more immersive 3D views. In fact, in addition to a traditional 2D panel, a 3D buildings viewer obtained using a virtual globe was implemented. The 2D viewer makes available to users all the typical WebGIS functionalities, while the 3D viewer provides a simultaneous visualisation of the urban context in a more realistic way. The advantage is quite evident: alongside the traditional Web-mapping solution, with a zenithal view, the user can navigate in the 3D representation. The second relevant value added in this application is the ability to navigate over time.

The proposed theme is of interest to both information and communication technology and cultural heritage audience. Moreover, that solution can be extended to all environmental and territorial areas taking advantage, in terms of analysis, from the visualisation of the temporal evolution of a phenomenon. The Web-mapping application presented has been implemented starting from open source available software in such a way to maximise the possibility to bend as much as possible the solution to our needs without having to develop the application from scratch. Such a system allows users to view maps of Como from different time series, made available by the local State Archive, superimposed on the current city map to appreciate the evolution of the city over centuries.

The system can be considered a major step forward compared to current solutions for historical maps online. In Section 2, in which the state of the art is shown, some interesting implemented solutions are presented that have been regarded as a source of inspiration for the original solution proposed by the authors.

The system takes advantage from the interoperability of the components used for its development, and it is able to access interfaces compliant to Open Geospatial Consortium (OGC) and International Organization for Standardization (ISO) specifications. The ad hoc solution adopted for the moment for the modelling of the data, even if not yet optimised with respect to de jure standards, is based on shapefiles, which represent a de facto standard for many GIS applications. In any case, further developments will consider the extension to other data models, mentioned in Section 3 of this article.

The final aim of the system is to create a multi-dimensional platform on the Web where users find historical information and updated geographical data of the city of interest, in order to navigate and analyse the evolution of the city in time.

Data used for our specific application are described in Section 4, while Section 5 states the technology and the data model adopted, both at server and client side. Section 6 shows the implemented functionalities, and conclusions are summarised in Section 7.

2. Background

Historical maps are a dependable source of information regarding past city planning. To analyse their content, the most practical method is the direct comparison with the present ones, for instance, by overlaying them. In recent years, many historical archives and libraries started to publish their cartographic collections on Internet. Several projects have been carried out by Italian State Archives in collaboration with universities, like the ‘Atl@nte dei Catasti Storici e delle Carte Topografiche della Lombardia’ (Atlas of historical cadastral and topographic maps), developed by the Politecnico di Milano with the State Archive of Milan (Oreni et al. 2010), and a collection of Venice Cadastral maps created by IUAV University and the State Archive of Venice (Contò et al. 2009). Other examples have been realised by the National Library of Scotland, which has georeferenced Scottish Town Plans to deliver a georeferenced map in a customised interface and created a mash-up (Henrie 2009), and by the publishing company Cartography Associates, which has created a Web geo-catalogue service, the David Rumsey Map Collection, with over 30,000 maps and images online.

The University of California, Los Angeles (UCLA), with the University of Southern California (USC) and the City University of New York (CUNY), has developed a digital research and educational platform for exploring, learning about and interacting with the layered histories of cities, called HyperCities (Presner 2010). This platform allows users to go back in time in order to create, narrate and explore the historical layers of city spaces and tell stories in an interactive, hypermedia environment.

In these projects, early maps have been georeferenced, in order to superimpose them on current cartography. Furthermore, these projects attest the high level of development of free and open source tools for Web publishing historical maps.

In recent years, advances in geoinformation technologies, such as 3D virtual environments, 3D analytical visualisation and 3D formats for data sharing, offer a large spectrum of new possibilities for communication of ideas and discussion of design alternatives (Isikdag and Zlatanova 2010). 3D visualisation, in fact, would give several advantages, for example, a completeness in the representation of geographic data, a more realistic representation with a greater expressive power, and more effective reading and analysis (Takase et al. 2008).

The easiest way to three-dimensionally displaying georeferenced maps is given by the so-called ‘virtual globes’, which allow users to explore the Earth in three dimensions while streaming satellite imagery, digital elevation model, and other geographic data from the Internet (Schultz et al. 2008). Referring to virtual globes, we can distinguish between closed and open source solutions. The most popular closed source technologies are Google Earth, Bing Maps 3D and Ovi Maps 3D. Those virtual environments have made 3D visualisation of urban fabric known and accessible for everyone. In the open source world, we can recall ossimPlanet, gvSIG3D, osgEarth, Norkart Virtual Globe and NASA World Wind (Walker and Kalberer 2010).

In our project, it has been decided to adopt an open source solution because of code openness and hence the possibility of customising it in the content and functionality. Among the possible technologies, NASA World Wind has been chosen. Its characteristic of being written in Java makes it platform independent and accessible by a simple Web browser like an Applet or a Java Web Start Application (Brovelli and Zamboni 2012). Furthermore, its mature state of development and the community support have been taken into account.

At the moment, there are no other examples of historical cadastral 4D Web-mapping systems, showing the evolution of the city. The solution proposed by the authors is an evolution of Web C.A.R.T.E., a system that enhances the accessibility of the cartographic heritage of State Archive of Como, implementing Web geoservices for cataloguing and visualisation (Brovelli et al. 2012). The improvement presented here and the whole system are based on standards; therefore, it can be reused for any historical archive. The originality of the system consists in providing the synchronous 3D/4D visualisation of the evolution of the city contextualised by using imagery and current and historical maps: the data model and the system were designed and implemented by the authors. The major complexity was the design of the system and, after the analysis of the partial available solutions (characteristics and performances), its implementation.

Moreover, due to the high level of detail that can be required by such a kind of analysis, we opted for a solution allowing us to use a detailed digital terrain model (DTM) instead of the global ones, made available by the proprietary virtual globes. The 2-m-resolution LiDAR DTM used by the system (with an accuracy of 40 cm) ensures a more realistic navigation within the model.

We can summarise saying that the general aim of our work was to create a new advanced viewer based on the available ISO and OGC standard interfaces, enhancing the mashing up via geoweb services of all maps useful for contextualisation, improving the immersivity by means of a virtual globe and a detailed DTM, allowing also the time navigation.

3. 3D city modelling and Web visualisation

One of the most important characteristics of every component in a geosystem for ensuring inter-operability is its conformity to standards. Information about city objects has therefore to be given in a standardised way so that they can be useful for different applications, such as environmental simulations, urban planning and disaster management.

A particular standardised format, based on XML, has been developed for virtual 3D city model representation and exchange, called CityGML. Based on Geography Markup Language (GML), it has become an OGC international standard. The aim of CityGML is to reach a common definition and understanding of the basic entities, attributes and relationships within a 3D city model (Kolbe 2009). In particular, CityGML represents four different aspects of virtual city models: semantics, geometry, topology and appearance.

All objects can be represented in five different Levels Of Detail (LOD), where an object become more and more detailed with increasing LOD regarding both geometry and thematic differentiation:

• LOD0: Regional model: the coarsest level. Essentially, it is a 2½D digital terrain model.

• LOD1: City model: block model, without any roof structure.

• LOD2: City/Site model: roof structure and larger building installations, such as balconies and stairs, optional structures.

• LOD3: Site model: architectural model with detailed wall and roof structures, doors, windows, and so on.

• LOD4: Interior model: completes the previous model by adding interior structures, such as rooms, stairs and furniture.

Those different LODs facilitate efficient visualisation and analysis: in the same dataset, an object can be represented simultaneously in different LODs, enabling the visualisation with different degrees of resolution.

CityGML is complementary to 3D computer graphic standards, such as X3D, VRML and COLLADA, and geovisualisation standards, such as KML. All these formats are based on the standard XML markup language, with the advantage of being both human and machine readable and readily usable for Internet transmission.

Once created, 3D city models can be visualised on the Web through OGC Web3D Services (W3DS). This specification is currently in draft status and not yet adopted by the OGC, but it has already been implemented in some projects, such as in the Spatial Data Infrastructure (SDI) for the city of Heidelberg in Germany (Basanow et al. 2008). The W3DS delivers 3D scenes of city or landscape models over the Web as VRML, X3D, GeoVRML or similar formats. The service is used not only for producing static scenes but also for requesting data, in order to stream it to the client which implements a more dynamic visualisation.

A similar approach has been adopted for a project that investigates the possibility of generating a 3D city model from OpenStreetMap (OSM) free geo-data (Over et al. 2010).

Concerning our project, at the moment we have opted for a simpler data model, which is not a de jure standard but, being described by a shapefile, can be considered a de facto standard. The main reason for that choice was the detail for data available, which are very rough: buildings are simply represented by block model, corresponding at most to the LOD1. Moreover, the emphasis at this first step was the implementation of the whole system leaving the choice of the optimised model as second step. The model will be described in Section 5. Further improvement will include the adoption of the CityGML model.

4. Available maps

From the beginning of 2010, the State Archive of Como has started a high-resolution digitisation of its different cadastral series:

• Theresian cadastral maps (1718–1722): this cadastre came into force from 1760, under Empress Maria Theresa. The map scale is 1:2000.

• Lombardo-Veneto cadastre (1854–1858): started in 1854. It has a scale of 1:2000 (with some attachments at 1:1000 or 1:500 for details).

• Updates for the Cadastre of Buildings of the Italian Kingdom (1898) at scale 1:2000.

• 1905 maps belonging to Land Cadastre (this is a Cadastre available only for some municipalities and, among them, Como) at scale 1:2000.

Besides their artistic value, these maps describe with significant accuracy the territory they represent, turning out to be precious instruments for scholars and professionals working on this area (e.g. for urban planning or restoration plans).

In order to display the old maps within a WebGIS system, they have been georeferenced and warped in a current reference system. Before starting with the georeferencing operation, a pre-processing step was performed. Most of the cadastral maps, in fact, are available divided in sheets, generally not perfectly matching. Therefore, after various tests, a choice has been done: instead of separately georeferencing each sheet, a complete map for each series has been obtained by unifying the map sheets in a proper way. The georeferencing procedure has then been applied to the mosaicked map. The main reason of that is due to the fact that the procedure is based on the search of homologous points between the ancient map and a current one and the subsequent warping. The scarcity of homologous points in map sheets that represent small portions of territory (example because peripheral) permits a poor results warping only.

Once the complete digital cadastral maps for Como were created, it was possible to proceed to the georeferencing procedure. Among the available interpolation methods (e.g. polynomial, rational functions and thin plate spline), the one that gives the best statistical result, that is, the polynomial, was chosen (Brovelli and Minghini 2012). The reference system in which maps are projected is the WGS84/UTM zone 32N.

An example of the georeferenced map of Como belonging to the Theresian cadastre is depicted in Figure 1.

Figure 1. Georeferenced Theresian map of the walled city of Como.

Ancient maps were portrayed in the WebGIS together with current geospatial information, in order to better contextualise their content. Due to its expressivity and high resolution, we decided to use for this purpose an orthophoto provided as a WMS trough the National Geoportal by the Italian Ministry of Environment, Land and Sea. This orthophoto was created by the orthorectification of aerial photos taken in 2006. It covers the whole national territory and gives us a detailed view of the area in the chosen reference system, with a spatial resolution of 0.5 m.

Moreover, to represent the built environment, the current digital map of the town of Como at the scale 1:2000 has been adopted, while the information on the different buildings has been drawn from master plans of Como town of 1975.

5. Data model and technologies adopted

As mentioned in Section 2, different data models are available for describing cities. In our case, as the emphasis of the project was on the creation of the Web application for both 2D and multi-dimensional visualisation of the city evolution, we have decided to adopt a very simple (but performing) data model, postponing to a second phase the use of standards such as those mentioned above. An ad hoc data model, based on a shapefile format and a ‘time-driven’ approach, has thus been developed. Following that model, if an object (e.g. a building) changes in shape or dimensions, the showed geometry changes too. The evolution in time is described associating to each polygon two columns in the attribute table of the shapefile: one with the ‘construction date’ and the other with the ‘demolition date’. This means that to every change a new geometry is associated, while the previous one dies. All the partial geometries that logically represent a unique building have the same identifier number (ID) assigned, in order to keep trace of the geometry modification of each construction.

Despite its originality, this approach is similar to the one adopted in CityGML specification: in the AbstractBuilding class the function of the building (e.g. residential, public, industry, etc.), the usage, the class (e.g. habitation, sanitation, administration, etc.), the year of construction, the year of demolition, the roof type, the measured height and the number and individual heights of the storeys above and below ground are in fact specified. The attributes we use are the equivalent of the two CityGML parameters yearOfConstruction and yearOfDemolition.

Moreover, also some proprietary GISs, for example, ArcGIS 10, enable time support specifying a Start Time Field and an End Time Field in the database associated to the geometry in order to define the feature life cycle. Therefore, the model adopted is very similar to other available solutions. The specification of time is standardised, compliant with the standard ISO 8601:2004 Data elements and interchange formats – Information interchange – Representation of dates and times.

Furthermore, also the mean height of buildings (if available) is stored as an attribute of the shapefile so that extruded models can be generated.

The application has been developed using free and open source solutions, and both server and client components have been customised. At the server side, MapServer has been used, while the main core of the client is based on OpenLayers and World Wind. Figure 2 represents the architecture of the proposed system.

Figure 2. The architecture of the system.

MapServer is one of the most spread open source Web mapping tools. Originally developed by the University of Minnesota (UMN) ForNet project in cooperation with NASA, and the Minnesota Department of Natural Resources (MNDNR), it is currently a project of the Open Source Geospatial Foundation (OSGeo). UMN MapServer is Free Software licensed with a MIT license, a free software license originating at the Massachusetts Institute of Technology (MIT).

MapServer supports numerous OGC standards, allowing users to publish and consume services and data in an application neutral implementation manner. In particular, MapServer can support WMS (Web Map Service) as server or client, WFS (Web Feature Service) as server or client, WCS (Web Coverage Service) as server and SOS (Sensor Observation Service) as server.

Concerning the application described in this article, MapServer has allowed us to publish both the historical maps and the building digital map as WMS services.

In principle, a WMS service allows users to integrate time-sensitive data. Specifically, MapServer provides support to interpret the TIME parameter, defined into WMS OGC specification, and to transform the resulting values into appropriate requests.

To be consistent with the server side and also for the client side, free and open source solutions have been adopted. A useful general comparison between those Web-mapping clients is available in Internet (Carillo 2012). The implemented geoportal is rather complex because it allows both the traditional 2D display, typical of WebGISs, and a multi-dimensional display through a virtual globe. The system is operated with two specific panels for each viewer, which can be moreover synchronised.

In the following two subsections, the technologies used in the project are described.

5.1. 2D visualisation

In order to make available a client providing users excellent experience and interactivity, a Rich Internet Application, based on HTML and JavaScript, was implemented.

Regarding the 2D visualisation, different libraries have been adopted: OpenLayers (version 2.11), GeoExt (version 1.0) and Ext JS (version 3.4.0). OpenLayers is an object-oriented JavaScript library for displaying maps in a browser, and it is free from server-side dependencies. It implements a JavaScript API for building rich Web-based geographical applications. As a framework, OpenLayers is intended to separate map tools from map data so that all the tools can operate on all the data sources. Furthermore, OpenLayers implements industry-standards methods for geographic data access, such as the OGC WMS and WFS protocols. OpenLayers is completely free and released under a BSD-style license (2-clause BSD License, also known as FreeBSD). GeoExt is a JavaScript library that extends Ext JS, a rich library of Web User Interface widgets and helper classes for building Internet applications. It integrates OpenLayers as a mapping library with Ext JS framework. GeoExt provides a suite of customisable widgets and data handling support that makes it easy to build applications for viewing, editing and styling geospatial data. GeoExt is released under the BSD license, while Ext JS is available with an open source license compatible with the GNU GPL license v3 only for open source applications. By means of such pieces of software, an intuitive and interactive geo-browser can be created.

5.2. 3D visualisation

A NASA World Wind Java (WWJ) Applet has been used for the implementation of the 3D viewer. NASA World Wind is a free and open source virtual globe written in Java. These two characteristics mean respectively that it allows developers to create their own interactive visualisations of 3D geographical data directly modifying (if necessary) the code (openness) and that it is platform independent. Moreover, NASA World Wind is a mature technology, developed by NASA staff and open source community developers. It was released in 2004 under the NASA Open Source Agreement, and it was ported to Java in 2007.

Using WWJ Software Development Kit, a pool of components to visualise 3D geographic information within applications and applets are available. Figure 3 shows the World Wind architecture. The applications or applets use World Wind by placing one or more WorldWindow objects in the user interface. This component provides the 3D geographical context for the application information and behaviours. In addition to WorldWindow, five main interfaces are present:

• Globe, which represents the shape of the planet and the terrain;

• Layer, which applies imagery, shapes or other geospatial information to the Globe;

• Model, which aggregates Globe and Layers;

• View, which interactively controls the user's view of the Model, via the InputHandler; and

• SceneController, which controls the rendering of the Model and the timing of the rendering. It associates a Model with a View.

Typically, an application associates a Globe and different Layers with a Model, and then the Model is passed to a SceneController that displays the virtual globe with its layers in a WorldWindow, with an interactive View. By default, the platform makes directly available a collection of pre-configured classes to project satellite images (Landsat7, Blue Marble, USGS orthophotos, etc.) and Digital Elevation Models (SRTM, ASTER, USGS NED, etc.) on the Earth as WMSs dynamically served by NASA and United States Geological Survey (USGS).

Figure 3. The major World Wind interfaces (image taken from http://goworldwind.org).

In the WWJ Applet, local orthophoto and DTM are used, that is, the national orthophoto with pixel resolution of 0.5 m and the Lombardy LiDAR DTM with ground resolution of 2×2 m.

The complete freedom of customisation makes the platform suitable for scientific applications, having the possibility of controlling horizontal (with the texture) and vertical (with the DTM) components’ quality and accuracy. In fact, it is possible to project on the globe whichever georeferenced image taken from an OGC compliant WMS server and to import whichever digital elevation model owned by the user by means of special WMSs.

At the server side, the most appropriate format to store 3D domains of irregular shape is the vector format. Such an information can then be made available from a WFS server or simply from a Web server by using a standard format currently used (e.g. KML, X3D, GML or shapefile).

In our project, the 3D model has been built up from a shapefile connected to an alphanumeric database containing, for each volume, the construction period of the building and other features, like the extrusion height of the polygons. The data loader is implemented directly within WWJ.

6. Implemented functionalities

Adopting the dataset and the technologies described in the previous sections, a WebGIS has been developed, available at http://historicalmaps.como.polimi.it.

The system was built starting from the requirements and considering the suggestions provided by the experts of the State Archive.

6.1. 2D/3D geoportal basic functionalities

As already said, in the HTML page of the Web site two panels have been created: one for the 2D map panel and one for the virtual globe. The interface of the portal is shown in Figure 4.

Figure 4. Interface of the developed WebGIS.

On the left side, there is a ‘classical’ map panel (2D frame), while on the right side, the virtual globe created with NASA World Wind (3D frame). The usual navigation tools have been inserted in the 2D Map Panel toolbar, such as the zoom in/out, zoom box, zoom to the maximum extent, pan and previous/next view.

The southern part of the page is divided into different tabs to manage several functionalities. The first tab (Raster Layer Tree) contains the list of the following raster layers published in the WebGIS:

• orthophoto (2006), served as WMS from the National Geoportal. It is the base layer;

• Carta Catasto Teresiano – Theresian Cadastre (1722);

• Carta Catasto Lombardo-Veneto – Lombardo-Veneto Cadastre (1858);

• update of the Lombardo-Veneto Cadastre (1873); and

• Carta Catasto Terreni – New Lands Cadastre (1905).

Acting on this layer tree, it is possible to turn on/off the early maps overlapped to the current orthophoto of Como, simultaneously on the 2D and the 3D frame. The order of layers listed in the tree determines how layers are drawn on a map.

By dragging layer names in the layer tree, it is possible, moreover, to change layer order in the map panel, setting top or bottom layers.

The second tab (Vector Layer Tree) is related to the list of the following vector layers:

• buildings under conservation programme, identified from a 1970's general urban plan;

• buildings historical classification, identified from a 1970's general urban plan; and

• ancient parishes of the XVIII century.

As in the previous raster layer tree, it is possible to modify the layer order by dragging layer names in the tree.

Those layers come from different classifications of the same dataset, stored in the Web server as a shapefile. In the 2D panel, the corresponding maps are provided by a WMS server and thematised according to the building construction period (‘Buildings Historical Classification’ layer) or to their ancient parish membership (‘Ancient Parishes’ layer). The ‘Buildings Historical Classification’ layer, instead, is provided as a WMS layer supporting also temporal requests in order to show only those buildings that fulfil the requested time span. Buildings under conservation programme are highlighted in the ‘Buildings under Conservation Programme’ layer.

Once a vector layer is turned on, its legend appears in the Legend tab. The content of the legend tab changes dynamically according to turned-on layers. The two legends available are that of the ‘Ancient Parishes’ of the city and that of ‘Buildings Historical Classification’, in which different colours correspond to different classes (e.g. grey for buildings built up before 1600, blue for those built up between 1600 and 1760 and so on).

The following tab contains the transparency settings: for each layer turned on, except for the base one, a transparency bar appears to let the user change the transparency settings of each layer. This useful capability helps users to understand city evolution in time. Figure 5 shows the 1858 map superimposed on the orthophoto with transparency set at 30%.

Figure 5. Lombardo-Veneto Cadastre map superimposed on the orthophoto.

The WWJ Applet panel offers all the basic features to get the information about the location on the virtual globe (latitude and longitude), the elevation (in metres) from the ground and the altitude of the view (in kilometres), the scale and the compass.

Besides those basic tools, in the WWJ Applet Tools and Time Display Management tab, users have a set of functionalities to act on the applet: it is possible to turn on/off the city model, to thematise the extruded buildings by colour or height and to change the view of the virtual globe.

6.2. 3D data analysis

Within the geoportal, by means of two sliders positioned below the two panels, it is possible to filter the buildings according to different criteria. With the first slider users can visualise all the existing buildings at a certain year, selecting a year with the cursor from nowadays back to 1600.

To perform this filtering in the 2D case, it is not possible to use a standard WMS with time support service because the year selected by the user has to be compared with two values for each building (the year of construction and the year of demolition) to check if this year belongs to the building existence interval.

However, using MapServer, this comparison can be done by a run-time substitution of filter variables inserted in the configuration file (the mapfile).

The second slider enable the visualisation of only those buildings built up during a certain time span chosen by the user from the year 1600 to the year 2000.

In this second case, the WMS time capability is enough to request the filtered map to the server.

An example is visible in Figure 6.

Figure 6. Example of selection for the construction period criterion. Only those buildings built up from 1610 to 1991 are shown in the two frames.

6.3. Synchronisation

Since the two panels are built up with completely independent technologies, it is necessary to develop a set of JavaScript and Java functions to let the two panels communicate. First of all, a JavaScript function has been inserted in the raster layer tree code to register the node status: if a map layer is turned on/off in the OpenLayers panel, then it is turned on/off also on the virtual globe.

The biggest effort has been done for the synchronisation of the location: if the synchronisation is enabled, the user can see the same portion of the Earth both in the planar view (2D) and in the perspective view (3D).

The synchronisation module registers every change applied with the navigation tool and passes the bounding box coordinates to a specific function for co-ordinates transformation. While the 2D panel makes use of planar co-ordinates, in fact, the virtual globe works with geographic ones, so it is necessary to transform on the fly the co-ordinates extent from planar (North, East) to latitude and longitude. This transformation is done with a JavaScript function that makes use of the traditional Gauss-Krüger formulas to pass from UTM WGS84 32N zone to WGS84 geographic co-ordinates. Having the view extent in the WGS84 reference system, it is possible to set the virtual globe view on the desired location.

7. Conclusions

In recent years, new Internet GIS tools have been developed to allow researchers, professionals but also common people a wider and more efficient consumption of geographic information. The presented prototype falls in this field: we started implementing the 2D visualisation through a WebGIS of ancient digitised maps belonging to historical archives; then, to better contextualise this information, we overlap those layers to actual orthophotos and maps. In fact, besides the artistic aspect, maps describe with great accuracy the status of the territory; therefore, they are a valuable analysis tool for many users, for instance, scholars or professionals working on a specific area (e.g. for urban planning and restoration projects). The overlapping task required a rigorous georeferencing and warping of the ancient maps, done by exploiting homologous pairs on the maps themselves and on a current map, taken as reference.

The webGIS has evolved into a more sophisticated system, called Web C.A.R.T.E., capable of navigating large raster data in a georeferenced framework and easily retrieving their metadata.

Working with the experts of Archives and ancient maps, we realised that for scholars, historians and professionals, a great improvement in the use of these maps is the availability of the third dimension and possibly also the fourth, that is, the temporal evolution of the city.

At the same time, in addition to the traditional 2D display of geographical data, virtual globes were implemented in order to extend the perception of the area of interest and to give a more realistic representation of the digital Earth. These tools are becoming familiar even to non-expert users in geographical data browsing. This leads to the need of revising the way data are presented in WebGISs, which will be no longer only 2D. Virtual globes are the tools we decided to exploit in order to improve our system. A novel Web-mapping multi-frame and multi-dimensional framework has been designed and implemented. The application is built customising and adding new functionalities in popular and well-known free and open source software packages. It is worth mentioning that the creation of such a system is not simply a trivial integration of existing packages but, on the contrary, before we conceived the system we verified which pieces of software could be reused and which were to be written from scratch. In our geoportal, a 3D panel, synchronised with the traditional 2D, helps users in data browsing and visualising both in space and time domain.

Being absolutely extensible, the geoportal can be used for many other environmental and territorial applications. Next step, to make it completely interoperable, will be the implementation of a specific plugin for reading a standard city model (as, for instance, CityGML). Moreover, processing functionalities based on WPS (Web Processing Service) will enrich the geoportal, proving also tools for analysing the visualised data.