Thank you very much for the introduction. Now I‘d like to talk about the project Colonia3D. It is the result of the joint effort of four
The structure of this talk is as follows: First , I‘d like to introduce the project results as well as its goals and challenges. I then will give an overview about the reconstruction process, followed by a brief overver of optimization techniques we applied for the real-time rendering. Then I introduce the concept and visualization techniques for the exhibition at the romano-germanic museum. Finally, I‘ll conclude my talk.
Let me start with the introduction to the results and challenges of our project.
These pictures show the remains of the roman culture that are still visibile in todays city. Our goal was to enable to reconstruct Cologne in 200 B.C. in a way that it can be explorated interactively in a virtual environment Let me start with a video footage of our results.
So to summarize, we were up to hit the following goals: We want to create an interactive virtual 3D reconstruction of the ancient Roman Cologne that can be presented as part of an exhibition at the romanic german museum.
Challenges: Another challenge was the implementation of the interactive visualization , since theres no existing framework for this prupose. Requirements visitor: create a sensation he could not get via the internet in a sourrounding
Let me now focus on the virtual 3D reconstruction [CLICK] For example, the detailed lion head on this slide was modeled using sculpturing tools.
To illustrate this process: [CLICK] First the archeologist create precise 2D drawings of the reconstruction which are then referenced [CLICK] in a 3D coordinates system by the designers. [CLICK] Based on that, they create a 3D model which is then exported to Collada and send to the rendering team. The main goal was to use the same model for off-line and on-line (interactive) render
Certain buildings, in particular at the rhein-front, where modelled at high quaility. This slide show a screen shot of the real-time renderer. Just to give you a proper comparison: [CLICK] this picture show a real reconstrcuted romanic column from Xanten. [CLICK] this picture show a virtual 3D reconstruction of the same item. One can see the details. [CLICK]
Besides visualization, buildings of such detail can also be used as a basis for creating solid input models for a 3D printer which output is depicted here. [CLICK]
Further, they can be used as background scenery for movie productions. For example for a student production that a famous speech in front of a the main temple.
But there a part of ancient city for which they amount of archeological data is quite small. There for the 3D reconstruction inherently contains two levels-of-detail. There are around 15 building reconstructed with a high level of detail, such as this villa depicted in front. [CLICK] For most other buildings, such as the insulae depicted in the back [CLICK], models remain in a abstract fashion.
So, to give you an overview: the reconstruction of the anchient Cologne comprises roughly 100 Buildings. [CLICK] [CLICK] [CLICK] [CLICK] [CLICK] [CLICK] As well as the tombs in the in the south and north of the anchient city.
Precomputed global illumination for real-time rendering
8 Colors Change quite often over time
Designer are up to excellent quality, since they spend a lot of work in the modells
The complete modell could not be edited in 3DS max, thus was in its form not suitable for real-time rendering. [Click]
Paradigm: Hardware over algorithms: Trend war absehbar – reduces the cost of developing out-of-core strategies
We use S3 texture compression with a ratio of 8:1 – this of cause comes at the price of visual quality.
Of course, the lossy S3 texture compression of the texture atlases comes at a price: [CLICK] The visual quality, as you can see on the examples of the shadows. This was a hard hit forthe designers, but it enables us the compression ratio of 8:1 [CLICK]
Okay, now we‘re able the render the complete 3D model in real-time. This is the foundation for the museums application which i like to describe now. [Click]
Mouse + key board vs. Space-mouse and touch table
We choose a client-server approach for the museum application.
Which i‘ll explain in greater detail in the following.
The objective of this mode is to enable the exploration of the complete model and points-of-interest, so called hot-spots.
Object highlighting Colors mapped dynamically, wheter object is disabled (grey), active orange, or highlighted
Also, communicate different reconstruction possibilities.
One further feature of the frameworks was to included in the concept is the display of multiple stages of an excavation and ist related findings.
When in comparision mode, the user can scroll panorama of current cologne, The projections show two stacked panorama – Todays cologne on top and acncient version (bottom) The user can also use the space mouse to control the panoramic camera.
So how did we implement this functionality? The solution is quite simple, both panoramic images are texture-mapped onto a cylinder which is place in the 3D virtual scene. „Panoramic Camera“ with base-line compensation
Different users: old vs. Young Experienced vs. unexperiecend
With respect to 3D we integrated the following approaches in order to support the user.
We use the bullet physics engine for collosion detection prevent the user to ran into walls or buildings.This requires an explcite collosion model
Parabolic Camera Path (slow acceleration at the start of the path and damping at the end)
Now to the conclusion of my talk.
So, what we have learned so far: First of all, a distributed and interdisciplinary project is very costly, not only in terms of money, but w.r.t. time,
Further, the reconstruction process was not as linear as we expected: (back and forth between designer and archeologist) (Changes in color, details, even complete reconstructions)
Then, since the application should run 24/7 the testing and debugging should start asearly a possible. (Different users) Problem: Speed of virtual camera Documentation for the museums pe Rsonal
Te Tuning of physics parameters
Of cause, there are a number of ideas for future work. And, of cause get founded.
So to conclude:
Finally I‘d like to thank the team of numerous reserachers and student workers which often completely volunteer their work to the project. Without them, this project would problably have never finished. It is nice to realize to have an impact by bridging the gap between science and people.
This concludes my talk and I hope you liked it. I‘m happy to answer any question you may have on this topic
Colonia3D (EuroMed 2010)
Communication of Digital Cultural
Heritage in Public Spaces by the
Example of Roman Cologne
Matthias Trapp, Amir Semmo, Rafael Pokorski, Claus-Daniel Herrmann,
Jürgen Döllner, Michael Eichhorn, Michael, Michael Heinzelmann
2. 3D RECONSTRUCTION
3. INTERACTIVE RENDERING
4. EXHIBITION AT ROMANO-GERMANIC MUSEUM
GOALS & CHALLENGES
Virtual reconstruction using polygonal 3D models
Real-time rendering of 3D reconstruction
Concept for romano-germanic museum (RGM)
Meet requirements of archeologists
Meet requirements of museums visitors
Implementation of interactive visualization
How to create an “experience”?
3 visualization scenarios
How to display meta data?
Integrated vs. multiple views
How to support the user?
Make use of modern hardware
SUPPORTING THE USER IN 3D
Avoid „getting lost“ situations
Support easy exploration of sites
Explicit collision model for physics engine
Constraint orbiter camera
Automatic camera path generation
Dedicated „Home“ scenario
Computed parabolic camera path
After 4 year development and 5 months up-time:
Distributed, interdisciplinary development is costly
3D reconstruction: multiple iterations required
Add scale elements
Test and tune the application and setup off-site early
Acquiring robust hardware
Press coverage & publications
Success can depend on political issues
POTENTIAL FOR FUTURE WORK
Make it better
Link to the tangible assets of museum
Include high detail findings
Conduct user evaluation
Indoor visualization and exploration
Make it available
Accessible via web using WPVS
Providing the framework for other cities
Licensing the model to the community
Interactive museum application of a complete city
Now permanent exhibition at the RGM
Extensible real-time visualization framework
Potential for future work
Open for collaborations
Official Press Release: 3.12.2010 in Cologne
Special thanks to the team
Johannes Bäuerlein, Henner von Hesberg, Fabian
Kampa, Simone Lamberts, Janine Lehmann,
Christian Leuenhagen, Stefan Maass, Silvio
Mölter, Sebastian Heilmann, Annette Paetz,
Christian Rademann, Christopher von Bronsart,
Michael Hoffmann, Solveig Böhl, Yu-Chung Chen,
Su Hyung Han, Enrichetta Minuzzi, Bierschenk
Martin, Sascha Blättgen, Omar Gerado Chacòn
Yela, Benjamin Jager, Sven Aufenvenne, Martin
Kramer, Benjamin Richter, and Leif Rumbke.