2): In the first step the cameras inner geometry and its position and .... Figure 6: The front right perspective view of a statue's head in Luxor temple, Luxor, Egypt ...
MONUMENT PRESENTATION USING DIGITAL ARCHITECTURAL PHOTOGRAMMETRY Klaus Hanke, Associate Professor Institute of Geodesy, University of Innsbruck, Austria Mostafa Abdel-Bary Ebrahim, M. Sc. Assiut University, Egypt, Doctoral Student at Innsbruck University, Austria
KEYWORDS:
Architectural Photogrammetry, Archaeological Photogrammetry, Digital Photogrammetry, Close-Range Photogrammetry, Digital Projector.
Monuments,
Documentation,
ABSTRACT Digital photogrammetric methods have made their way also into architectural applications which is traditionally a rather conservative branch. Beside the programs and techniques that have been adopted from aerial photogrammetry mainly low-cost systems are used for architectural and archaeological applications. New strategies for data acquisition, restitution and presentation of the digital data are under development.
INTRODUCTION
DIGITAL METHODS
Digital methods spread in all fields of photogrammetry today. Easy handling, the comfortable and clear processing with widely accessible computer equipment and all the advantage connected to this (low cost of hardand software) have lead to its ultimate success. The availability of digital images especially for close-range applications with Photo-CD and digital cameras as well as the valuable amount of data shows this field to be very suitable for digital photogrammetric solutions.
Digital photogrammetry adds a variety of new features to that development. Origin of the digital images can either be digital cameras or scanned analogue photographs. For the latter task the Photo-CD developed by Kodak has proofed to be very useful. It dispenses the photogrammetrist from the necessity of owning an professional scanner and delivers a high amount of geometrical stability and accuracy of the scanned photos [Hanke,1994 and Hanke et al.,1995]. Beside this, the compact disk is an excellent media for storing data without loss over long time. Moreover this method is very well suited for application of already existing or archived metric photos up to a format of 4 x 5 inch with resolutions up to 4096 x 6144 pixels. Digital cameras of comparable quality are on the contrary rather expensive or are provided with a small image matrix or low resolution. A very strong development in this direction is within sight.
The classic results of architectural photogrammetry are facade maps, profiles and perspectives. In addition to that, 3D grids, surface and volume models in connection with CAD systems come out of the spatial photogrammetric restitution. As in other surveying branches the latter have the advantage of easy continuation of data processing and administration without restriction to single analog maps and plans. Rectification of metric photographs is also a classic method in photogrammetry. Originally there were only photos of planes which have been rectified, later there appeared also approaches to perform differential rectification of photos showing curved or general surfaces of objects [Voizikis, 1983]. These approaches came to their limits as they are restricted to surfaces of the form z = f ( x , y ) meaning that the third coordinate is only a smooth mathematical function of the x and y coordinates which is very rare in architectural applications. Architectural objects are generally not to describe in that easy way because they have ledges, alcoves and niches, hidden parts and especially old buildings have often rather irregular surfaces. Therefore, generally a really spatial (3D) reconstruction of buildings has to be performed also because a trivial partitioning into single facades (e.g. with curved surfaces) is not always possible.
Regarding the photogrammetric restitution the advantage of digital photogrammetric techniques become visible. Very expensive analytical plotters become replaced by common available PCs or workstations. Digital methods of image matching and image correlation support the user in performing the inner and exterior orientation of the images for further processing. Automatic processing of the image’s contents really is a challenge in architectural applications. Other than in aerial photogrammetry the content we look for is not always clear and very often not or very hard to describe by algorithms (even by knowledge based systems). Because of the extension in depth of the architectural objects with their prominent or set-off sometimes hidden parts of the surface every smoothing algorithm must fail. No object point can be used as an approximation for its neighbor as we are used to do it in topographical applications. Between them could lie a difference in depth of several
meters. The solution for this can only be seen in a continuos interaction between user and software. Once be set to a line that has to be evaluated the software will be able to follow it on its way. The information extracted from this line can be e.g. used to restitute parts of an object using spatial spline curves [Forkert, 1993]. Another successive way is to use sketches of the required facade elements to extract these lines from the images [Steilein,1995]. The progress in digital methods for architectural photogrammetry may lie in this sense rather in an interactive and partly automatic support of the user than in an indifferent full automation of the process.
operation. The acceptance of respective lectures in architectural photogrammetry made the realization of a number of smaller and larger project in Austria and in other countries possible. The participation at international scientific initiatives allowed to verify and advance own developments.
THE CONCEPT OF THE ”DIGITAL PROJECTOR”
On the contrary to the conventional ”rectification” of metric photographs the ”Digital Projector’s” concept is a strict object oriented three-dimensional restitution of the whole object without almost any conditions and limitations. The approach is based on a reversion of the situation during exposure. Figure (1) shows the scheme of the approach.
At the Institute of Geodesy, Innsbruck University the authors made an effort to contribute to this development. The current contact to architects, conservationists and archaeologists led to a fertile and constructive co-
The concept of the ”digital projector” approach is on one hand a new rigorous digital photogrammetric tool and on the other hand takes into account the demand for an object oriented, three-dimensional reconstruction of monuments.
digital or analog camera (data acquisition)
scanner, Photo-CD (digitalisation of films)
bundel adjustment (point cloud, orientation)
image enhancement (preparation for projection)
CAD system (frame- and surface model)
3D rendering (digital projection)
video, Photo-CD, printer, plotter ...
Figure 1: Data flow of digital architectural photogrammetry The process of reconstruction of a monument to get a virtual computer model happens in 3 steps (Fig. 2): In the first step the cameras inner geometry and its position and orientation during exposure as well as the building’s characteristic framework of outlines and faces is to be reconstructed mathematically. To get a homogenous solution of the total object these elements are computed in a so-called ”bundle adjustment” where a probable result of the whole measurement system will be created. Beside of this way the restitution of the outlines may also come from an analytical plotters result. In the next step within a CAD environment this framework will be reviewed and if necessary completed where also additional measurements (photos, tape, theodolit, etc.) can be added. After that the 3D model will be closed defining faces between the structure lines and will be investigated for leakage performing a rendering process. The face
model that arise from that step will be used as a kind of ”projection screen” and can be of very different degree of details. In the third step happens the actual reprojection of the photos [Hanke, Ebrahim,1996]. Similar to a slide projector some selected photos are projected to the surface model of the monument. The selection is done regarding the visibility and its direction of projection. The ”Digital Projectors” own the same inner geometry and relative position as the related cameras. These values come out of the bundle adjustment can be different for all cameras. The result of this process is a complete 3D computer model of the object from which
b) CAD-edited frame model
a) Frame model from bundle adjustment
Figure 2: Steps creating a virtual computer model
d) Digital projection onto the model surface
c) Surface model (”projection screen”)
e.g. orthoimages of single facades (Fig. 3 and 4), perspective views (Fig 5 and 6), arbitrary animations or interactive applications can be derived. The approach is so general that not only architectural applications but also archaeological documentation, medical or forensic applications are possible. As the restitution is strict object oriented there is no problem combining very different photos of the same object. The use of close shots for interesting details and overview shots to include neighboring objects with less resolution of projection is also possible without problems and shows the flexibility of the approach. It is also possible to combine the inner and outer parts of a building. Some of the projectors just have to be situated in the inner space and do the virtual projection onto the inner walls.
CONCLUSION There is a fundamental change in regarding the results of a surveying to overcome all of us. Following this development there will be not only single plans and maps drawn anymore, moreover the measured data will be administrated in structured information systems that do not anticipate any further use of this data. Translating this to the task of architectural photogrammetry that means not only to measure and restitute single facades but also to record the monument as a whole. Plans of facades, perspective views or profiles and even virtual reality animations will be only derivative products of a coherent prepared database of this monument. The way of using the derived data is not anticipated through this database.
The kind of presentation (i.e. perspective or orthogonal view, global or any interesting detail) will be decided by the end user of the data. The combination of photogrammetric, geometric and arbitrary expert’s data (building material, style, owner, arts, builder etc.) leads to the creation of a Monument Information System (MIS). In combination with Digital Terrain Models (DTM) and geocoded (rectified) aerial photographs the imbedding of the building into its natural environment is possible. The transition to digital urban models is from now on just a matter of scale, i.e. of the possibility for generalization of the data model. Conservationists and archaeologists incline rather to maintain existing and proved methods, as it is according to their profession. The immense advantage and chances of digital systems and results will cause a change of views in architectural photogrammetry. Virtual worlds are nowadays accepted and used already by planning architects to visualize new projects through animations.Its victory in documentation and representation of existing or already lost cultural heritage is still pending.
ACKNOWLEDGEMENT The shown projects are partly from the international CIPA test project ”Otto-Wagner-Pavillion, Karlsplatz Vienna” or have been supported by the Austrian Government, students work during different lectures at the University of Innsbruck, Austria and diverse sponsorship of local authorities.
Figure 3: Orthoimage of a historic railway station in Voels near Innsbruck, Austria
Figure 4: The front orthoimage of Ramesses III temple in Luxor, Egypt
Figure 5: 3D-Computermodel of medieval tower in Innsbruck, Austria
Figure 6: The front right perspective view of a statue’s head in Luxor temple, Luxor, Egypt
BIBLIOGRAPHY AND REFERENCES Forkert, G. (1993) ”Photogrammetric Object Reconstruction Using Free-Formed Spatial Curves”. In: Gruen/Kahmen (Eds.) Optical 3D Measurement Techniques II, Wichmann Verlag, Karlsruhe. Zürich 1993. Hanke, K. (1994), ”The Photo CD - A Source and Digital Memory for Photogrammetric Images.” IntercongressSymposium of ISPRS-Commission V, Melbourne, Australia. In: International Archives of Photogrammetry and Remote Sensing, Volume XXX, Part 5 Hanke, K., Weinold, Th. (1995), ”Using the Photo CD as a Source and Digital Memory for Photogrammetric Images - A Report on its Data Compression Method and the Geometrical Stability of the Transferred Images.” Proceedings of the St.Petersburg - Great Lakes Conference Digital Imaging and Remote Sensing 95, 25. 30. June 1995, St.Petersburg, Russia, GUS. SPIEPublications, Series P, Vol. 2646, Washington DC , USA
Hanke, K., Ebrahim, M. A.-B. (1996), ”A General Approach for Object Oriented 3D-Mapping in Digital Close Range Restitution.” In: International Archives of Photogrammetry and Remote Sensing, Volume XXXI, Part B5, ISPRS-Congress Vienna 1996. Streilein, A. (1995), ”Videogrammetry and CAAD for Architectural Restitution of Otto-Wagner-Pavillion in Vienna.” In: Grün/Kahmen (Eds.) : Optical 3DMeasurement Techniques III, Wichmann Vozikis, E. (1983), ”Digitally Controlled Differential Rectification of Mathematically Defined Surfaces.” Photogrammetria , Vol. 38 (1983), S 165-180. Waldhäusl, P. (1992), ”Defining the Future of Architectural Photogrammetry.” Invited Paper. In: International Archives of Photogrammetry and Remote Sensing, Volume XXX, Part B5, ISPRS-Congress Washington DC 1992
