Bob Martens
Vienna University of Technology

Before individual activities in the field of endoscopy are to be explained a brief insight into the surrounding thereof may prove meaningful. At the Department for Spatial Simulation endoscopy is not treated independently, but principally in connection with other simulation techniques such as the simulation of architectural spatial formations in full-scale, CAAD, stereoscopy and holography. The term SAAD (Simulation Aided Architectural Design) [The term "SAD" (Simulator Aided Design) was coined at the likewise named SAD-Laboratory at the Tampere University of Technology] refers to a combination of spatial simulation techniques. This aspect plays a major role at the Vienna University of Technology. The laboratory is situated in the main building of the Vienna University of Technology at Karlsplatz. The facilities available to the lab are directly adjacent to the centrally located 1st Courtyard and are accessible on the groundfloor via the courtyard through the main entrance. Further linkages are located in the basement and also lead directly to the stage area. The endoscopy area is also located there next to the experimental stage and the respective storage rooms. The experimental areaof the laboratory consisting on the ground floor is surrounded by a working gallery open vault (figure 1).

Figure 1:

Both from this gallery and from the working area on the ground floor one can watch any on-going experiments The experimental stage has a clear height of practically 9 meters and a working area of approx. 60 m2 The experimental floor and the ceiling resp., consists of adjustable girders and floor units making for alterations of the floor area in any desired manner. The panelling of the side walls consists of the same wood plates as the floor units, so that things like screwing, nailing can be accomplished without difficulties similar to work on a theatrical stage. At the half height of the room the removable banister of the working gallery becomes visible. The movable working platform has not only been designed for mere observation purposes. It is supposed to be used for experimental work, too, for suspending or positioning light-weight building members as well as for the operation of lighting fixtures in the busbars provided. Any lateral movements of this working stage are performed manually by a person on stage using a hand crank. The experimental floor consist of adjustable floor girders and panels making for any required height alterations. This floor-modulation-installation was already taken into account while construction work was under way, thus the later wall brackets were dimensioned appropriately to be able to bear the required loads and were equipped with holes every 17 cm.

Installation of the Lab Area for Endoscopy

The acquisition of a rigid endoscope in combination with a very inexpensive CCD-video camera in 1989 set the stage for the first steps in the field of endoscopy. A VHS-recorder with monitor was at our disposal as peripheral equipment. Two years later the recording procedure was expanded to an extent that "single frames" and "quicktime movies" now are transferrable unconditionally into the computer and can be further processed there. The complete equipment is designed for still pictures and simple sequences, as these mainly deal with the special field of interior design. Installation of mechanical equipment - of a camera rig in order to control motion sequences - is scheduled for the long run. First equipment was installed provisionally in a seminar room, even though the exposure table with background screen, the light fittings and the endoscope with peripheral equipment had to be installed newly for every recording set. Meanwhile a better solution for this situation was found by adaptation of an own lab area for endoscopy (figure 2).

Figure 2:

Considerable improvements were accomplished with the initial equipping of this area. A soft source (a so-called "windowlight" with 5000 W) is mounted shiftably on two rails over the exposure table with background screen. Sun is simulated with an additional light source (1250 W). Presently a highly sophisticated ball-bearing studio tripod is being used. The endoscope with a viewing angle of 50 degrees is mounted together with a CCD-camera positioned appropriately for picture taking. This miniature camera surely does not furnish top-quality pictures, its robustness, however, has proved very useful regarding the amount of handling it is subjected to by the students. So required we can also rely on a C-mount adapter and furthermore on a connection to the media center of the Vienna University of Technology. The equipment as described above has been in operation since the end of 1992.

Image Processing as Main Goal

The unpopularity of endoscopy in the representation of architecture can be traced back to problems in the peripheral area. When viewing an architectural model by endoscope, e.g. the definition of the picture is fine, but the image the viewer receives is hardly conveyable to others. In order to facilitate communication several experiments were made to connect the endoscope with peripheral recording media:
A: Photographic Technique

The endoscope is attached to the lens of of a reflex camera by means of a transitional ring. Only still pictures can be made.

Apart from transitional ring and spotlight, so required, hardly any additional investments are required. Comparing to video-pictures a much higher resolution is achieved.

Endoscopes are of low light intensity and require a lot of light. Working conditions when taking pictures seem rather unpleasant. While the photographer is selecting the picture section in the finder of the camera he is exposed to a considerable amount of heat depending on the amount of spotlights being used. Trial runs are required in order to define correct period of exposure this making for frequent interruptions caused by film development every time.

A1: Film-Technique
The endoscope is attached by means of transitional ring to the lens of a film camera.

As in (A).

As in (A). The lower degree of light sensitivity cannot be made up for by means of extended periods of exposure.

B: Video-Technique

The endoscope is connected with the video camera by means of an adapter. As to suitability of camera no general conclusions can be made: it has to be tested for every specific instance.

The result becomes visible directly on the monitor, i.e. any desired modifications can be immediately performed. The medium video is not principally limited to still pictures: movements and camera drives can be registered in various video formats (S-VHS, Betacam SP, Hi8, etc).

Up to approx. 1980 only Ólow-qualityÓ video-cameras were available in the semi-professional field. The CCD-camera has, however, done away with this problem to a large extent. If no editing equipment is available, the raw material has to be further processed somewhere else.

B1: Digitizing / Digital Video

A computer plant with image processing software has to be equipped with a digitizing card which enables processing of e.g. PAL- or NTSC video-signals. Some software-packages have so-called plug-ins which import the images without detours. Export possibilities of computer sequences on video require further investments regarding the hardware (e.g. Genlock-Interface).

As with video-technique

Relatively high investments necessary. A tremendously high memory requirement necessary when e.g. real-time-sequences are to be depicted (per second 25 with frame- and 50 with field-recording). Quality limitations, however, result anyhow by the PAL-video-system regarding picture resolution. Thus the video-picture has to be digitized sufficiently large, as subsequent "blow-up" is no solution (figure 3).

Figure 3:

B2: Still-Video

For completeness sake still-video has to be mentioned. Individual pictures are taken with a still-video camera and are stored on mini-disks. The pictures produced can be played on a monitor or may be processed in the computer. Any definite experience in connection with endoscopy as far as registered has not been published as of yet.

Storing in digital format by means of a small, handy camera.

Still-video pictures so far only provide of a mediocre resolution. Considering that endoscopic exposures are predominantly taken in laboratories a good individual picture could be rescued from the video tape, so required.

Both as far as digital video and still-video are concerned overlappings with photographic and film techniques can be determined. Real pictures taken photographically are achieved providing a considerably better definition by scanning, so required, and mounting with the digitized videoframe. For pragmatic reasons the AppleMacintosh-vicinity has been chosen for work at the Vienna University of Technology. Presently, it just does not work out that the entire working capacity is dedicated to computer work. Detailed considerations as to the advantages and disadvantages of various operating systems would be too lengthy for this paper. Photoshop - a registered trademark of Adobe Systems - is being used as image processing software. Amongst others, it is characterized by a vast range of differentiated tools, manipulation filters and plug-ins. The lack of definition in the digitized output-picture resulting from the hardware shortcomings (figure 4) was somewhat relieved by means of the filter "sharpen".

Figure 4: Figure 5:

Then the picture was completed as the terrain of the model taken proved too small. For this reason existing picture parts, such as the water surface, were duplicated. The subsequent addition of human figures (scanning in, scaling and positioning) could also be feasible. So-called "facet-filters" were also put to use. Lack of definition was thus softened even more by the effect achieved by painting in water colors. Finally the approx. 1.5 megabyte picture was transferred to the EDP-center for diascoping (figure 5) with a resolution of 2800 DPI [This equipment as presently configurated is designed only for Postscript documents.Apple Macintosh rather works with other formats e.g. Tiff and Pict. Luckily, we managed to proceed with printing in the usual fashion by making use of software "Interprint" by Intercon. The document is first translated into Postscript and then checked into the printer queue.].

Figure 6 is showing the same model, but viewed from a different site. The question here is how the reflection of the building in the water was achieved. The required building parts were selected with the tool "magic wand" and was copied into an additional document. After vertical rotation the reflected picture information was inserted in the original document and the reflection was accordingly distorted. By means of mounting control convergence was reduced to approx. 70%. Furthermore, the reflection in the water was manipulated for the selected water surface with the filter "shear" (figure 7).

Figure 6: Figure 7:

Expectations for the Near Future?

Principally, first encounters with the medium endoscopy should produce positive associations. This is to be interpreted that the demonstration is to involve potential possibilities rather than exclusively stressing its limits. Though 3D-computer simulations are presently in fashion and endoscopy is rather considered as "out" the short training time has to be particularly pointed out. The required scale models are being built anyway and only call for insignificant adjustments. Endoscopy is not overloaded with theory; students can tackle the subject in a "playful" manner.

The question is, how many endoscopes have been purchased for architectural purposes world-wide and how many of them are really still being used. We do not have any exact figures, those having merely become dust-catchers could be revived without too much effort. The reinstatement could prove simple in the low-cost range, as many universities and offices nowadays have efficient computer plants at their disposal. The variety is large starting with the simplest still picture to the pre-programmed computer controlled movement via camera rig. Furthermore, the mixture of "real" and produced reality (by means of endoscopy) must be considered. A renaissance of endoscopy could thus be envisaged at least for still pictures. As for the moving pictures the exposure and editing of sequences via computer presently is only to be realized with certain limitations. Despite compression procedures there is a tremendous storage requirement. What seems likely is that the increased use of multimedia will inevitably lead to advancements in the hard- and software area.

The combination of endoscopy and stereoscopy, referring the taking of stereoscopic endoscope pictures (both stills and animations) is a practically unexplored research subject. Also the combination with holography - the endoscope acting here as a Holographic Optical Element (HOE) - has so far not been investigated. Finally, the endoscopic optics represent a further field of investigation. The development of rigid endoscopes with high light intensity could prove very beneficial. With specific model simulations - particularly regarding indoor areas - lighting always poses a problem. If the companies supplying endoscopes invested vigorously in such advancements (even though medical applications are by far much more rewarding regarding commercial aspects), endoscopy could probably gain in interest as far as applications in the field of architecture are concerned.

Bob Martens