A method of transcribing archaeological sites from oblique aerial photographs

Journal of.4rchaeological

Science 1976,3, 391-394

A Method of Transcribing Archaeological Sites from Oblique Aerial Photographs Rog Palmer”

Introduction In the publication of the recent symposium on aerial reconnaissance for archaeology Irwin Scollar demonstrated two simple graphical methods for transferring information from oblique air photographs to maps (Scollar, 1975). One of these, the Mobius network method, relied on networks constructed on the air photograph and the map that used the same identifiable points on both. The principle of the Mobius network is that intersections so formed correspond on both photograph and map. Thus an archaeological site on an air photograph crossed by a network can be transferred to its relative position on the map network. Given a site on flat ground and with sufficient control points, a network can be constructed that allows the salient features of a site to be transferred to the map with fair accuracy. The problems caused by different ground height are difficult to overcome by conventional transcription, and good control points will always be important, but the following method of transcription eliminates the unnecessary lines of Scollar’s network and increases its accuracy. Method This method also relies on the Mobius principle, but with the addition that any line on an air photograph parallel to the horizon plane is itself of constant scale. It must be emphasized, however, that perspective on an oblique air photograph causes each of these lines, although of constant scale over their own length, to be of different scale to any other parallel line. It will often be found that the horizon plane of most hand-held oblique air photographs is parallel, or nearly parallel, to the frame of the photograph; although being a function of the position of the camera at the moment of exposure this is not always so, and it represents the vanishing line of the camera. Because it is important that this line is always established a method of dealing with excessive tilt is outlined further below. The basic method of transcription is shown in Figure 1. (1) Construct a line ab parallel to the horizon plane through two suitable points on the photograph, and another through the corresponding points AB on the map. (2) Set proportional dividers to match the two different scales. (3) Draw lines on the photograph from other control points, through significant points of the site, to intersect line ab. (4) Using proportional dividers, transfer the points of intersection on line ab to line AB and reconstruct the network on the map. “Department

of Archaeology, Downing 391

Street, Cambridge, England.

photograph

Map

“.

Figure 1. Illustrating the method of transcription of selected points of detail on the air photograph to the map via a network. To aid clarity only part of the site has been transcribed. (Based on a photograph by J. R. Boyden.)

Aerial

TRANSCRIPTIONS

FROM

AERIAL

PHOTOGRAPHS

393

(5) When sufficient intersections have been made through points of the site on the air photograph, the complete shape can be accurately transferred to the map. The advantage of this method over the “random” network described by Scollar is that it pinpoints selected detail of the site immediately and with maximum precision. Scale Selection Using a large scale map (a minimum scale of 1 : 10560 is necessary for reasonably accurate transcription) and all available air photographs, an extremely accurate plan can be made of an archaeological site. This may include, for example, pits, entrance complexes and kinks in ditches, any of which may help explain the functions of the site. But such a plan takes a long time to construct and the need for such intricate detail may be questioned, especially if the plan is made as a prelude to large scale excavation or is to be used in conjunction with a geophysical survey as both will amost certainly reveal features that did not show on air photographs or are of a more complex nature. My own work has suggested a map scale of 1 : 5000 as ideal for the transcription of most sites or complexes that are to be found in Britain. The ground area covered by the map can be sufficient to show any significant relationships, while the scale is large enough to show all but the most trivial details of a site, and the level of accuracy attainable is adequate for planning partial excavation and for morphological study. The pedantic plotting of sites at larger scales offers little useful information beyond that of a 1 : 5000 transcript, especially when it is considered that any plot from air photographs must be regarded as an interim statement (Hampton, 1975, p. 119) and that ground inspection is a necessary precursor to any useful description of the site. Correction of Tilt Lines of constant scale on an air photograph are perpendicular photograph, and for frames with excessive tilt this line can (Figure 2). (1) Define a square (ABCD) on the map whose corners photograph (abed). Due to the relatively large scale of most

to the principal line of the be determined as follows can be identified on the archaeological oblique air

Figure 2. Showing the construction necessary to determine the principal line for a tilted photograph.

394

R. PALMER

photographs the square on the map is easiest constructed by the use of a grid drawn on transparent film which can be moved over the map until any four identifiable points are located. The sides of the square can be at any angle to the frame of the print. (2) Join ab to meet dc produced, and similarly, join ad to meet bc produced. This is best done by parallel construction through a point e. (3) Join the intersections of these lines to give the vanishing line, or true horizon of the print. (4) Mark the principal point p of the print. This is the centre of the complete frame, frequently indicated by fiducial marks on the print edges. (5) Draw a perpendicular through p to the vanishing line. This is the principal line. (6) A suitable line of constant scale can now be drawn perpendicular to the principal line through identifiable points on the photograph, and can be used to scale off lines of a network in the manner described above. (Based on Kilford, 1970, p. 43-44.) Summary and Conclusions The network transcription method described above is simple to use, can be speedily carried out, and may be as accurate as the needs dictate. An archaeological site can be precisely located and drawn in correct proportion by the use of a few well placed lines. Combined with more critical and experienced interpretation the adoption of this practice for air photograph transcription for the growing number of county “Sites and Monuments” record maps, for example, would give greater value to such maps to those of us who are not just County Planners. The method may also be of some help in transcribing part of the vast backlog of air photographs currently awaiting processing, although with this in mind it is hoped to be able to offer a simple and rapid computer method in the near future which may overcome the problems of transcription-but not those of interpretation. Acknowledgements I am grateful to Collin Bowen and David R. Wilson for their comments on an earlier form of this note. References Hampton, J. (1975). The organization of aerial photography in Britain. In (D. R. Wilson, ed.), Aerial Reconnaissance for Archaeology, 118-125. London. Kilford, W. K. (1970). Elementary Air Survey. London. Scollar, I. (1975). Transformation of extreme oblique aerial photographs to maps or plans by conventional means or by computer. In (D. R. Wilson, ed.), op. cit., 52-59.