3D geo-informatics has entered the digital age, hesitantly in some areas, and rampantly in others. Google Earth and Microsoft Virtual Earth are household names. Although limited to landscapes and buildings envelopes, their massive digital geographic libraries are today the playground of millions of people and the generator of new forms of content and applications with tremendous impact perspectives. However, these pale in comparison to those that will be made possible as soon as urban digital libraries will be fully available and exploitable.

Therefore, the V-City project aims to research, develop and validate an innovative system integrating the latest advances in Computer Vision, 3D Modelling and Virtual Reality for the rapid and cost-effective reconstruction, visualisation and exploitation of complete, large-scale and interactive urban environments. The focus of the project on urban environments is not only made possible by the latest technological advances, but also justified. Urban environments represent one of the most important and valuable cultural heritage as acknowledged by the UNESCO.

This system will enable historians, architects or archaeologists to reconstruct from existing data, study, understand, preserve or document urban environments using an innovative interactive 3D user interface. This project will progress beyond the current state of the art in the field of large-scale geospatial libraries built from multi-source and multi-format architectural and cultural information.

It will also be an answer to concrete needs for a wide range of users as demonstrated by the commitment and the diversity of the end-user organisations involved in the V-City User Group. These will contribute to both the definition of the system and its validation on real-scale scenarios.

Description of target users and groups

Historians, architects and archaeologists

Description of the way to implement the initiative

Scientific & Technological Objectives

The V-City consortium provides the world-class and complementary competencies required to tackle the following scientific and technological challenges:

1. Processing of a new generation of geo-referenced images: multi-angle oblique pictures captured from the ground, with airborne systems operated by the consortium, or satellites. Filtering, perspective correction, tone homogenisation, and correlation of these images with other multi-source and multi-format geo-referenced data (LIDAR point clouds, cadastral data, DEMs...), extraction of building shapes and facade textures.
2. Automatic extraction of the topological and semantic information about the constructions (floors, doors, windows, style, shape of roof...). This work will be based on the latest research results published by the consortium at SIGGRAPH 07.
3. Reconstructing complete cities using procedural rules derived from the extracted topological and semantic information. This work will extend the recent research results published by the consortium at SIGGRAPH 08 and 06
4. Implementing techniques for optimising, storing and visualising online large scale geographic, architectural and environmental datasets in real-time, from global scale to building interiors. These techniques will also support non Euclidian information, such as textual information, topological and semantic tags, or exogenous data enriching the environment. This will extend the research results presented by the consortium at SIGGRAPH  07.
5. Implementing new multi-user interaction metaphors to intuitively browse, manipulate and interact with the displayed 3D urban datasets. This encompasses the research and development of a new kind of device, the interactive 3D map table, featuring a multi-touch multi-user table used to display and interact with the 3D reconstructed city.
6. Integrating these modules in a robust, seamless, distributed and time-critical framework that will serve as a backbone for all types of online urban activities.
7. Providing a V-City Software Developer Kit enabling third-parties to extend the system by implementing additional modules or gateways to other services.
8. Making the whole system run on entry level PCs or laptops, with the appropriate graphic board and hard drive.
9. Validating the impact of the proposed system on real scale end-user scenarios

Technology solution

Technological Background

The V-City project builds on the results of major European initiatives involving the consortium members:

The VirtualGeo system: VirtualGeo is one of the most-effective planet viewers. It has been developped in the frame of the European Virtual Planet (2001-2004) and CRIMSON (2005-2007) projects by CS with the support of CNR and CRS4. These projects have produced major advances in the field of massive geographic environment and large scale urban environments 3D visualisation. VirtualGeo is today sold by CS to professionals such as civil security organisations, aerospace industrials, or land planning and mapping agencies.     

The CityEngine system: this system is a revolutionary tool for the efficient creation of digital architecture. It has been created by PI and KUL members and improved in the EPOCH FP6 Network of Excellence. Recent research results published by these partners at SIGGRAPH 06, 07, and 08 enable today the launch of the V-City project.

BLOM's exclusive geo-referenced multi angle aerial imagery technology and digital library (tens millions of aerial images), opens radically new perspectives in terms of reconstruction of real urban environments.

 

Results

The V-City project researches and implements a complete and tightly integrated tool suite made of the following components:

 

The V-City Builder

The V-City Builder automatically reconstructs large-scale urban environments at high quality from input cadastral data, aerial oblique images and ground pictures. First, building data from multiple input sources is processed and a textured mass model is generated for each building. Second, topological and semantic information is automatically extracted and put into a compact procedural representation of the facades.

 

The V-City Server

The V-City server is the part of the V-City system in charge of transforming large scale representations of urban environments into good quality low-bitrate multiresolution representations suitable for streaming and rendering. It automatically optimises, compresses and stores the geographical and architectural data published by the V-City Builder and streams in realtime these data over a network connection with the V-City Explorer.

 

The V-City Explorer

The V-City Explorer is a next-generation globe viewer based on Diginext's future VirtualGeo3 system offering unprecedented rendering capabilities. It can display massive and extremely detailed urban environments, seamlessly integrated into the 3D landscape. Thanks to the BlockMap technology, developed expressly for this project by CNR and CRS4, the V-City Explorer can navigate through massive cities, complete with picture-perfect facades and architectural details, at breakneck speeds.
Although globe viewers are becoming more and more common, cities are rarely shown in 3D, let alone with accurate facades and architectural details. With the V-City Explorer, you can fly through a city in 3D, experiencing it in unprecedented detail, and visualising the city evolution along the time. The Explorer can even take you inside buildings, where you can discover the same hallways, rooms, and furniture that existed or still exists in real life!
The V-City Explorer is not just a simple browser, but a powerful tool for architects, historians, land planners, and other professionals. It includes sophisticated querying, analysis, or measurement tools and image capture. Professionals can analyze virtual cities with sufficient precision to draw accurate conclusions without having to leave the tool to consult other sources. Multi-platform support means that the V-City Explorer works on Windows and Linux, further encouraging collaboration.

 

The V-City Map Table

The V-City map table is intended as a multitouch, multiuser, stereoscopic interface to allow both the visualization and interaction with the massive urban data that are displayed by the V-City Explorer. It had to cope with specific constraints such as the real-time display of 3D data,the intuitive and efficient manipulation and edition of the content, as well as to provide collaborative visualization capabilities. Two technologies have been identified, which are the immersive 3D visualization equipment and multitouch tactile input. They have been brought together into the V-City map table.

Main results, benefits and impacts

The project has entered its third and final year, and already many advances beyond the state of the art that it developed have received attention from both industry and the research communities. GeoConcept, a leading European Geographic Information editor, has for instance decided in January 2011 to adopt VirtualGeo 3 and its SDK, developed in the frame of V-City by DIGINEXT in collaboration with CNR and CRS4 with the support of BLOM ASA, as the rendering engine at the heart of its future flagship product GeoConcept 7, which is distributed worldwide to professionals in the fields of culture and education, urban planning and public works, homeland security, rescue and defence.

The V-City Builder is another important procedural urban modelling tool developed in V-City by Procedural Inc in collaboration with Katholieke Universiteit Leuven. Procedural integrates and markets this result in their CityEngine product. The adoption of this tool is promising with users such as Dreamworks, Pixar, nVidia, IBM, Microsoft, Boeing, Navteq ... Procedural was even selected as a Winner of the Red Herring's Global 100 award, honouring the year's most promising private technology ventures from around the world, for its work on procedural modelling.

Finally, the VCity Map table developed by Immersion is one of the most important results of the project. It provides for the first time a multitouch, multiuser surface displaying a stereoscopic view of the environment adapted to each user. A new step toward the holographic table featured in the Avatar movie...

The project was invited to present its results in conferences such as SIGGRAPH 2010, where it was selected in the Emerging Technology panel, the ACM conference on 3D Web technology and Eurographics, as well as scientific journals such as Computer Graphics Forum, and the EURASIP journal on Image and Video Processing. International press also acknowledged the significant progresses made by the V-City project with articles and interviews in Euronews, the New Scientist, or La Tribune, among others.

Primarily focused on the Cultural Heritage sector, the V-City system is expected to benefit other domains that all share the need for affordable, easy to build and interactive virtual mock-ups of detailed urban environments.

Return on investment

Return on investment: Not applicable / Not available

Lessons learnt

This field will be completed by the submitter when the lessons learnt have been identified and understood.

Scope: International