POMPEII

The photomosaics of the facades of the buildings of Insula VI, 1 along via Consolare and vicolo di Narciso showed that the objective of producing photo-realistic and dimensionally accurate images of historical standing structures could be achieved.  Additional field trials were performed during the summer of 2005 and spring and summer of 2006 to further improve the process and increase efficiency:

Camera Alignment – Photogrammetric research had showed that the axis of the lens of the camera must be normal to the plane of the subject in order to produce orthographic photographs.  The camera lens must be aimed directly at the facade and must be level to produce acceptable images for the photomosaics.  Field experience demonstrated that this was best accomplished with two people and with a bubble level mounted on the camera.  One person sighted the camera toward the facade and the other used the level to assess if the lens was horizontal.

Photographic Distance – A survey of the streets in Pompeii indicated that the width between opposing facades generally varied between four and eight meters.  Several experiments were performed to evaluate the field of view of the camera at these distances.  Test photographs were taken of a leveling rod placed horizontally against a flat wall at incremental distances.  The photos clearly showed how much of the wall had been recorded (see figure).  Using the Nikon D-100 camera with lens set at a focal length of 18mm (35 mm film camera equivalent of 27 mm focal length), the following horizontal fields of view were visible:

Distance            Measurement with            Measurement with
From Wall          Camera Horizontal            Camera Vertical

8m                               9.9m                                  6.5m
7m                               8.9m                                  5.9m
6m                               7.3m                                  4.7m
5m                               6.2m                                  4.2m
4m                               5.0m                                  3.4m

The farther the camera was from the wall, the larger the photographed area.  However, at the fixed resolution of 300 ppi, each image also had a different scale.

Scale – Each test photograph was studied to calculate its scale by comparing the actual length of the leveling rod shown in the image with its length measured on the image.  Those taken five meters from the wall with a lens focal length of 18mm, and then displayed at a resolution of 300 ppi were determined to be very close to a scale of 1:25.  This scale produced photographs with excellent detail that required minimal resizing.[1]

Camera Orientation – The field distance tests also confirmed that it is preferable to take photographs in portrait format (tall rather than wide) by holding or mounting the camera vertically.  This orientation captures as much of the building height as possible, many times avoiding the necessity of taking additional upper level photographs.  The tests showed that at a distance of five meters, the maximum height that can be recorded is 6.2 meters, but only if the camera is elevated 3.1 meters.  If the camera is hand held or mounted on a standard tripod, only about four to five meters of the facade can actually be photographed.

Photographic Interval – At a photographic distance of five meters and with the camera mounted vertically, the maximum width that can be recorded is 4.2 meters.  If photographs are taken every two meters along the plane of the facade, each will overlap the next by over 25 percent on each side as is shown in the figure.  This overlap is adequate, but field experience showed that a closer interval sometimes helps when creating the photomosaic.  More importantly, additional photographs should be taken in the center of openings (doors and windows) and directly in front of columns, fountains and other features that might present perspective challenges when assembling the images.

Focal Length – Photographs taken at a distance of five meters produced images at the desired scale with a lens focal length of 18mm.  Additional experiments were performed to determine if other lens settings are more appropriate.  Test photos taken at a distance of six meters with a focal length of 22mm and seven meters with a focal length of 26mm produced the same sized images as the ones taken at five meters distance and 18mm lens setting.  Identical photographs can therefore be taken at different distances if the correct focal length is known and used.

Photographs taken with wide-angle lens settings are generally more susceptible to distortion.  Photographs taken further away from a facade, but with a longer focal length, reduce lens distortion.  The limiting factor is the width of the street.

Control Backgrounds - The control background made of the facade of Insula VI, 1 along vicolo di Narciso worked well as a guide for assembling the photographs for the photomosaic.  However, the process of measuring and drawing offset dimensions every meter along the street was a slow process.  Also, establishing the horizontal line from which the measurements were taken was not only time consuming, but also required that strings be affixed to the facade.  A more efficient measurement strategy was needed.

Total Station surveying instruments combine a high-power telescope with an electronic theodolite, an electronic distance meter (EDM) and a computer into a single apparatus (see figure).  They are used for a variety of mapping and construction tasks in order to determine the three-dimensional coordinates of points on remote objects and features accurately.  The telescope is used to sight a reflecting prism mounted on a rod that is placed directly over a particular point.  The theodolite measures the angles from the Total Station to the prism, and the EDM uses the travel time of a reflected beam of infrared light to determine the distance.  The coordinates of the point are then computed and displayed in real time.

Technical advances now enable shorter distances to be determined without using a prism by measuring the travel time of a laser light beam that is directly reflected from a subject.  A Reflectorless Total Station allows one person to locate remote points and to determine their coordinates quickly.  This technology seemed appropriate for creating control backgrounds.

A Leica TCR 307 Reflectorless Total Station was used for field trials during the summer of 2005 to produce control backgrounds of several facades.[2] The building frontages were recorded from survey station marker points located in the street with known coordinates.  The telescope was used to identify the location of points on the facade such as along the ground line, sides and tops of walls, doorways, windows, lintels, holes and other diagnostic features.  The Total Station measured the angle and distance of each identified point, and calculated its three-dimensional coordinates.  The survey method was efficient and did not require any physical contact with the buildings.

At the completion of the survey, the data were downloaded from the memory of the Total Station to a laptop computer.  The coordinates were then imported into a computer-aided drafting (CAD) application at a scale of 1:25.  The appropriate points (facade outline, doors, windows, etc.) were then connected with a polyline tool to create an orthographic line drawing of the facade.  After a scale and notes were added, the drawing was exported as a bitmap image at its native resolution of 72 ppi.  The bitmap image was then imported into Photoshop where it was resampled at 300 ppi to create the final control background, which formed the base layer for the photomosaic.  Any photographs incorporated into the Photoshop document were automatically imported at the same resolution of 300 ppi.

Color Control – The use of reference photographs to adjust the color of the photomosaics was found to be somewhat subjective.  An independent and consistent color standard was needed.  A Kodak Q-13 Color Separation Guide (shown in sidebar on left) is often used as a reference when photographing artworks and artifacts. The card provides guidance for future color comparisons and reproduction.  Tests were conducted in the summer of 2005 to develop a technique to use a Kodak Q-13 card with the building exteriors.

Immediately before a facade was photographed, a Kodak Q-13 card was placed against the surface and a close-up photo taken in the same light in which the balance of the images would be taken.  After the photography was completed the photographs were transferred to a computer.  The photo of the color card was opened in Photoshop and its parameters adjusted until the screen image colors matched the Kodak Q-13 card.  The color settings of the image of the card were then used to adjust the color parameters of the photographs of the facade.  If the lighting conditions of all of the images are similar when photographed, the colors can be satisfactorily adjusted with a photograph of the color reference.

The Kodak Q-13 color reference makes it possible to adjust the color of a photomosaic on a particular computer and monitor.  If the photomosaic file is viewed on a different monitor, the color may not appear the same and adjustment will be necessary.  Printed images may not look the same as they appear on the monitor of the computer from which they were obtained, and will also require adjustment.  The color card should therefore be left on one of the Photoshop layers for future color comparisons.  The layer can be turned off when it is not needed.