Among the most enduring survivals from the past, in many parts of the world, is a wealth of ‘rock art’ created by pecking, carving or rubbing the surface of boulders and exposed rock faces to create petroglyphs – images of people, animals, fantastic creatures and symbols. These frequently survive in environments where painted pictures could not realistically withstand the ravages of time.
The reasons why these images (or ‘motifs’) were created, who made them and in some cases precisely how they were made, remain among the great mysteries of the past. For the great majority of the material, too, interpretation of the original meaning of the petroglyphs and painted pictographs is fraught with difficulty.
Rock art across the world has proved highly durable but there are now significant threats from climate change and land-use adaptation. There are even risks from increasing tourism, promoted – paradoxically – by the very research and publications that have made the presence and fascinations of the phenomenon more widely known. One tends to think of rock art as something that will last forever but in the face of these dangers there is a real possibility that some of the more fragile or easily accessible examples will soon be lost to us. Even those in our remote highlands may be at risk.
One of these remote locations lies high in the Syunik mountains of southern Armenia where the members of the Ughtasar Rock Art Project have in the years since 2009 identified and recorded over a thousand rock art panels 3,300 metres above sea level in the crater (caldera) of a long extinct volcano. Upper parts of the volcano are snowbound for up to nine months every year but in the short weeks of summer the crater supports a rich flora of grassland dotted with short-lived carpets of wild flowers. This remote and scenically spectacular location, marked by the now-shattered rock streams of once-molten lava flows, is also the haunt of wild animals including bears, wolves, foxes and wild boar.
“One tends to think of rock art as something that will last forever but there is a real possibility that some of the more fragile or easily accessible examples will soon be lost to us”
The Ughtasar petroglyphs include both abstract symbols and figurative motifs depicting long-horned goats, leopards and bears as well as humans and even wheeled carts and a primitive plough yoked to a pair of oxen. Early human activity within the crater is demonstrated by scatterings of stone tools and arrowheads made of obsidian, a naturally occurring black volcanic glass. The specific character of these finds suggests summertime use of this high pasture during and after the early copper age, which began in this area around 6,000 to 5,000 years ago. There is also some pottery from later prehistory and even a few petroglyphs that may date from the medieval period.
The exceptionally harsh conditions that prevail within the crater for much of the year mean that some of the petroglyphs, which are ‘pecked’ only one to five millimetres deep into the smooth and shiny rock surface, have been damaged by localised flaking as a result of the repeated freezing and thawing of moisture that has seeped under the surface crust of the rocks, in some cases removing parts of otherwise surviving petroglyphs. Any kind of walking on the petroglyphs, whether by rock art researchers or other curious visitors, can disastrously hasten the losses through this kind of degradation. Sadly, a few of the petroglyphs have been damaged by the thoughtless addition of modern graffiti.
Before any serious attempt can be made to ‘interpret’ the rock art at Ughtasar a detailed factual record must first be made. Within rock art studies that has traditionally been done by taking rubbings or, more recently, by tracing the outlines of the motifs onto flexible drafting film fixed temporarily to the rock surface. Both processes, of course, can cause minor damage to any of the petroglyphs that are already in a fragile state of preservation. In addition, while these ‘contact’ methods of recording can work reasonably well on flat surfaces, they are quite difficult to apply at Ughtasar where many of the motifs have been pecked into curving surfaces that have to be viewed, and then recorded, from a variety of angles and in a various lighting conditions before their true form and interrelationships can be properly assessed.
These difficulties led the members of the Ughtasar Project to realise at an early stage that more flexible ‘non-contact’ methods, and in particular various types of photography, would need to form a fundamental part of the recording process in the short periods of high-altitude exploration and survey that take place each summer once the team members have come together from their studies or year-round occupations in various parts of Armenia, the UK and elsewhere.
“The key to obtaining accurate models lies in the systematic and speedy collection of sharp, high-resolution, digital photographs during the hectic few days spent on site each summer”
Crucially for the success of the recording work, rapid advances in computer technology and digital photography have now made it possible to use various forms of photogrammetry to create viewable 3D models of the petroglyphs in their immediate surroundings. These photographs and digital models in turn allow the motifs to be closely studied, precisely measured and comprehensively described, revolutionising the way in which we can record the source material on site and subsequently make it available (admittedly at one remove) to anyone in the world who possesses a reasonably modest computer and a secure internet connection. Skype will permit real-time discussion between viewers, while Sketchfab and the internet will allow simultaneous access to the 3D models for researchers and rock art enthusiasts worldwide, few of whom will ever be able to visit Ughtasar in person.
The models are created from multiple digital photographs, using mathematical algorithms to calculate the position from which each image was taken. Once these locations have been established the software can calculate the relative position of each pixel within the overlapping photographic frames. Relatively high-spec computers with advanced 3D graphics cards and large amounts of random access memory are needed for this processing work, but the ultimate key to obtaining accurate and clear models lies in the systematic and speedy collection of sharp, high-resolution, digital photographs during the sometimes hectic few days spent on site each summer.
During the life of the project a wide range of photographic recording methods and equipment have been used and modified in response to practical experiments undertaken both at Ughtasar and at the Landscape Research Centre in the UK. In recent years these have relied heavily on the Agisoft Corporation’s multi-image photogrammetric software known as PhotoScan Pro, the availability of which influenced the evolving recording strategy as it has been modified and refined over the years.
Before the widespread use of PhotoScan and other multi-image visualisation techniques in archaeology, full-frame digital cameras had been used at Ughtasar for the selective recording of particularly important rock-art panels. But in the unstable rock streams this proved difficult and time-consuming, prompting a perceived need to work more rapidly and at increasing levels of detail. From 2012 onwards this led to a search for higher resolution cameras that nevertheless fell within the Project’s relatively limited financial resources.
One of the most encouraging advances came in 2016 when a SIGMA dp0 Quattro camera was purchased for research and practical application in various aspects of 3D imaging in the UK. The camera’s innovative three-layer Foveon sensor produces images that are quite different in their colour rendering and sharpness compared with the sensors found in most other digital cameras. In addition, the initial cost was less than £1,000 including the optional optical viewfinder.
“The dp0 Quattro when handled correctly produces images of extraordinary quality, making it ideal for use at Ughtasar where the petroglyphs are shallow”
In some ways the SIGMA dp0 Quattro, with its unusual design and fairly limited specification, is much more like an old film camera in the sense that for this kind of computer-based application it should be used at an ISO equivalent of 100 or 200 so as to reduce the already minimal risk of digital ‘noise’ in the resulting images. The camera also lacks the anti-shake mechanism that is now almost routine in recent mirrorless DSLRs. Despite this, however, the dp0 Quattro when handled correctly produces images of extraordinary quality, making it ideal for use at Ughtasar where the petroglyphs are shallow and the texture and colour of the shiny rock surfaces are highly uniform.
First tests during a brief visit to the site in July 2016 produced such good results that the camera was left with the research team for use in that summer’s fieldwork, part of which involved the recording of 19 rock art panels that were photographed in detail with the dp0 Quattro generously loaned to the Project each summer by English Heritage.
Once the July site trials were completed the processing of the images was undertaken in PhotoScan using Apple and Windows-based computers. This produced a remarkable set of 3D models, demonstrating that the SIGMA camera, used correctly in good natural lighting, could generate images and resulting models that matched and sometimes exceeded the quality returned by camera/lens combinations that were three or four times more expensive to acquire.
The PhotoScan process relies on the collection of multiple photographs that completely cover the target area, the individual images overlapping one another by a minimum of 60 per cent in all directions and the viewpoint being approximately vertical to the surface being recorded. The images are first checked for sharpness, any exposure variations are minimised and the format is converted from RAW to TIFF.
The resulting files are then imported into PhotoScan and processed using a standard workflow, which includes the calculation of image alignment from which the camera positions are precisely located. This information is then used to create a high-resolution point cloud that provides the basis for a mesh model that uses triangles of variable sizes to map the shape of the 3D target. Once the geometric model has been created in this way, small fragments of the original photographs are mapped to produce a high quality ‘texture’ of tone and colour that seamlessly covers the target area from whichever direction the model is currently being viewed. Scale bars are used in the model as the basis for setting the scale of the resulting model.
Once the 3D models have been created in PhotoScan they can be exported to Sketchfab web-site which is dedicated to the dynamic display of 3D models in internet browsers or for incorporation in a blog post or other web page. Sketchfab not only provides 3D display and navigation of the models, but also supports functions such as the application of textual annotations and image enhancements, including the addition of ‘digital lighting’ so as to use highlight and shadow to emphasise the shape of the parent rock or the clarity of the pecked motifs.
One of the research objectives of the Ughtasar Project has been to identify and develop methodologies, which can be shared with colleagues in countries like Armenia where expensive high-tech solutions are in most cases unaffordable. The Agisoft Corporation has supported the Project through the provision of a free PhotoScan license; more generally it provides academic licenses for research purposes at 15 per cent of the standard price for commercial uses. In addition, the creators of Sketchfab have granted the Project a free licence for the professional version of their software as part of the organisation’s international initiative to support heritage research projects and museums.
As a final observation on the rock art recording at Ughtasar, the SIGMA dp0 Quattro, although a little difficult to use in severe weather or poor lighting conditions, has demonstrated the merits for this kind of work of its innovative sensor/lens combination, securing superb results at a much cheaper cost than the use of more expensive full-frame DSLRs.
The Ughtasar Rock Art Research Project is sponsored by the Landscape Research Centre in the UK and the Institute of Archaeology and Ethnography of the National Academy of Sciences of the Republic of Armenia under the directorship of Tina Walkling in the UK and Anna Khechoyan in Armenia. Professor Dominic Powlesland is the Director of the The Landscape Research Centre.