Aerial Photography with a Tethered Blimp
Survey at Kerkenes was initiated because it was thought that a plan of the extant remains could be obtained with relative speed and at comparatively little cost by the employment of remote balloon borne photography. Experimentation at Yarasli in 1991 had shown that this was possible (Summers and Summers 1992; Summers 1992). Balloon photography was a viable alternative to aerial photography from planes, for which it is still impossible to obtain permission in Turkey. At Kerkenes photography with a blimp was completed in only two seasons, 1993 and 94, with a few holes in the coverage being filled with a kite in 1997.

Equipment and Method
Equipment used at Kerkenes was unchanged from that described in detail in Summers 1992. Today we would use lighter equipment and perhaps groups of smaller balloons or, better still for such a large site, a motorised paraglider. We would also use narrow angle lenses because they give less distortion. Essentially we used a helium-filled advertising blimp to carry a sling containing an Olympus camera with an automatic winder and a model aircraft radio control to fire the shutter. On still days, about one third of the days at Kerkenes in the summer months, the blimp could be flown in a series of traverses at altitudes of around 300m. Operators can usually hear the shutter or rewind with each frame so that, with a little practice, good sets of overlapping pictures can be obtained. On really still days, and with some courage, the blimp could be flown at altitudes approaching 800m. using a strong kite line rather than a rope for the tether (because the maximum height attainable is largely limited by the weight of the tether). Targets on the ground, in our case marked with white lime and plotted with a total station, permitted rectification.

Results
Results from blimp photography include:
1 High level oblique pictures that are great aids to interpretation and are, at the same time, visually striking. Results are, however, somewhat hit and miss since the operator has to estimate both the direction and the angle of the camera when the shutter is fired.
2 High level vertical photographs that can be rectified and used for mapping visible features.
3 Low level vertical photographs that can be rectified and used for detailed interpretation and mapping of visible structures.

Aerial Photography from a Manned Hot Air Balloon
The advantages of flying the photographers over the site are self-evident. Restrictions are cost, availability and weather. Civil aviation authority is required in addition to the research permit from the Ministry of Culture.

Stereo Photography from the Turkish Directorate of Land Survey and Registration
The Directorate of Land Survey and Registration (“Tapu ve Kadastro”) has stereo aerial photographic coverage of much of Turkey. These photographs are, however, subject to very strict security control that effectively makes them unavailable for archaeological research. Other official bodies in Turkey also have photographic archives that are subject to the same controls. MNG Ltd, however, had the requisite security authorisation to obtain a set of photographs of the Kerkenes Dag from which they most kindly made three-dimensional digital map of the site. Production of this map was crucial to the development of the project. Completion of the Global Positioning System (GPS) survey in 2000 should make the earlier MNG map redundant, but the size of the GPS data files and the complexity of the GPS software mean that the simpler map is still extremely useful.

Photo Rectification
Aerial photographs need to be rectified before they can be used to compile maps. Failure of the camera support to hang vertically below the blimp, in such a way that the camera is precisely parallel to the ground, is the most important cause of distortion. Lenses also distort images, and the wider the angle of the lens the greater the distortion.

At Kerkenes we have relied on AERIAL software produced by John Haigh from the Department of Mathematics at the University of Bradford. This program is able to rectify single images provided that there are at least five known points on each picture. We have rectified actual photographs and, more usually, rectified features traced from photographs. Software and hardware are advancing so rapidly that there is not yet a set or standard procedure.

Our present target is to drape our aerial photographs, and the interpretative plans made from them, over GPS images.