Home Marcus_Abbott.html
About About.html
ContactContact.html
Case StudiesCasestudy.html
 
Stonehenge
In 2012 English Heritage funded the most detailed analysis of Stonehenge. ArcHeritage were commissioned to conduct this analysis, our research has revealed significant new discoveries about Stonehenge including many previously unrecorded prehistoric carvings.
The full story of the discoveries can be read in the Nov-Dec 2012 issue of British Archaeology issue number 127 The English Heritage research report can be downloaded here http://www.britisharchaeology.org/http://services.english-heritage.org.uk/ResearchReportsPdfs/032_2012web.pdfshapeimage_7_link_0shapeimage_7_link_1
In April 2012 I received 850 GB of data from English Heritage, the data was the first compete high resolution laser scan survey of Stonehenge. With a point density of 0.5mm, this data set had the potential to provide new scientific information about the stones at Stonehenge. As part of the original methodology I created for English Heritage I had agreed to explore the visualisation possibilities for this data and hopefully create a working methodology for future studies of heritage assets similar in nature to Stonehenge. An important factor to the success of this project was the team ArcHeritage had assembled to carry out the research. Hugo Anderson-Whymark a freelance lithics specialist and expert on Stonehenge and I formed the main research team, with the specialist consultancy of Mike Parker Pearson, Colin Richards and Rob Ixer, we had a team which consisted of some of the finest experts on Stonehenge and its geology. The experience of this team was crucial to meeting the overall research objectives, it allowed us to put all of our discoveries into context and provided a valid archaeological backdrop to our work.
When we started working on this project it really was a leap into the unknown, as far as we knew a project of this scale had not been attempted before, certainly not on such an iconic site like Stonehenge. Needless to say there was great expectation and a bit of trepidation attached to this project. We started out examining the data in mesh form, the 1mm laser scan survey had been completely meshed and provided us with 3D geometry to examine. The examination was achieved in a virtual environment using Cinema 4D. In a workflow which can be likened to scientific visualisation; we were able to treat these huge stones, some weighing up to 45tons, as if they were small artefacts, studying them as a lithic expert would study a piece of flint, a technique not unfamiliar to Hugo. In a virtual environment we were able to enhance the surface details of the stones by creating custom made textures and applying them to the mesh models. We also experimented with different lights and the properties of light in the virtual environment. These techniques allowed us to visualise the surface details of the stones that had been captured in the original survey.
The results of this stage of the project were quite astounding, individual tool marks made 5000 years ago by the original builders of Stonehenge could be seen, not just single marks, but also layers of tooling which gave us an insight into the processes of Neolithic stone-working.This evidence had survived five millennia of weathering and erosion, it was also present on virtually every stone at Stonehenge. In order to control this information and to create a useful research tool we built a GIS and created a full catalogue of Stonehenge, every stone now has 1:40 scale elevation images the catalogue forms an atlas of Stonehenge. Plotting every tool mark onto this atlas provided us with contextual information about the patterning and distribution of the different Neolithic stone-working techniques. It became clear that groups of stones have different working practices, which might lead one to make suggestions about the chronology of the stones and the order in which they were created. Evidence seems to suggest that the inner Trilithon stones are worked in a different way to the outer ring, this might indicate that this inner construction came first. It is important to note that this project has generated so much new information that any conclusive ideas as to the effect this will have on the overall interpretation at Stonehenge, has not yet had time to develop and doubtless to say eminent prehistorians will be debating the significance of this new data for quite sometime.
Turning our attention to the known rock art at Stonehenge we set about locating and evaluating every documented carving on the site. This part of our research evaluated all carvings including those which were disputed. Our research has discounted many supposed carvings as being either natural geological features or reinterpreted them as areas of different neolithic stone working techniques. These features when seen in the context of the all the stone working we identified, appear to be just different areas of stone working; with examination onsite it is clear that these areas are more obvious to the observer and thus had previously been incorrectly identified as carvings.
Our next task was to locate the Bronze Age axe carvings on stones numbered 53, 4 and 3. These stones had a total of 41 previously known carvings. The 1mm mesh data which had been so successful for the identification of Neolith stone working practices, was proving to be an insufficient resolution to define the Bronze Age carvings. Whilst we could see the carvings in the data we noted that there seemed to be a data signature which looked as if there might be further discoveries to be made. This tantalising glimpse of new discoveries led us to experiment with numerous techniques in an attempt to visualise what we believed to be further prehistoric artwork.
We used a virtual studio to experiment with Reflectance Transformation Imaging (RTI). This studio consisted of 290 lights arranged in a hemispheric pattern about a point of origin. We positioned a full 0.5mm resolution mesh of the target stone surface at the point of origin and filmed the surface of the mesh as we engaged and extinguished each light in sequence. The interplay of light and shadow on the surface of the mesh would provide definition to the carvings and highlight any new discoveries. The 290 images generated from this technique were mathematically synthesised into an RTI file which can be viewed using a viewer allowing the end user to control the position of the light. Information on this technique can be found here at Cultural Heritage Imaging This technique confirmed that there was previously unrecorded Bronze Age rock art on the stones at Stonehenge. http://culturalheritageimaging.org/shapeimage_13_link_0
The RTI experiments also indicated that there might be even more discoveries, again we could see a data signature similar to the one we had previously observed in the 1mm mesh data. We decided to go back to the origin of the mesh and examine the original point cloud data. We used a piece of software called Pointools, which is designed for the visualisation of point cloud data. The key feature we utilised was the plane shading function. This function assigns a greyscale value to each point based on its distance from a plane derived from the camera viewing location. Essentially we created a greyscale band 7.5cm wide with a range from 0-100 white to black, this band was moved through the point cloud data at 1mm intervals. We moved the band 75 times so that each point would complete a full transformation and would have had all of the greyscale shades assigned to it. At each 1mm interval we rendered out a high quality image of the plane shaded data. These 75 images were combined to create an animation. When the animation was played back we could see rock art features fade in and out of view as the band moved through the data. Significant new discoveries were made using this technique, we plotted everything we identified onto our catalogue and were able to geo-reference the discoveries to 1mm accuracy on the British survey grid.
From the outset of this project it was always the intention to create transferable techniques so that this research would benefit not just Stonehenge, but a wider research community who can apply these techniques to their own specific sites of interest. With this in mind I ran a parallel experiment in data collection using photogrammetry. Originally the objective of this experiment was to capture higher resolution data than the 0.5mm provided, and use this as a method of evaluating exactly which features could be identified at specific resolutions. This was intended to guide our examination of the main laser scan data set, but as we completed some of the photogrammetric work, we began to realise that this technique was a valuable addition to the project. We used Agisoft Photoscan to capture data at a higher resolution than the 0.5mm laser scan data, varying from 300-200 microns in point density. This allowed us to build super high resolution mesh panels of areas we wanted to examine in more detail. This work concentrated on areas where we had previously observed prehistoric rock art. Armed with our new meshes we created a brand new visualisation technique to examine this dataset.
Luminance Lensing is a creative idea developed for the examination of high resolution mesh data derived from photogrammetry. I was interested in developing a low cost solution which could be employed on heritage projects that do not have the benefit of large research budgets. The technique was developed with a low cost workflow in mind and minimal software costs. Luminance Lensing works in Cinema 4D, it combines the success of plane shading and the functionality of realtime visualisation. The technique uses a material which is assigned to the mesh, the polygons which make up the mesh emit more or less light based on the interaction with a hypothetical 3D shape. The result is a realtime visualisation of the surface of the mesh on which rock art features glow or get dimmer than the surrounding mesh surface depending on various settings. Using this technique we were able to confirm the existence of the new discoveries and make some new ones. The findings of this project were always going to attract much attention, it was important to be able to backup our discoveries with a sound methodology. The creation of this new independent data set was crucial in providing a cross-reference for our discoveries. The fact that these carvings exist in two separate independent data sets allows us to be confident that what we had identified was a real feature and not an artifactual product of computer processing. I have no doubt that what we have discovered are extremely faint and heavily eroded carvings from the Bronze Age. Onsite the carvings are not visible to the naked eye but by using cutting edge technology we have been given a rare glimpse into the culture of a society that existed over three thousand years ago.
We have identified an amazing 72 previously unknown prehistoric carvings at Stonehenge, this doubles the number of this Bronze Age rock art in the entire country and brings the total at stonehenge to 115. We have scrutinised every square centimetre of every stone at Stonehenge and made over 700 discoveries, which help to increase our understanding of the construction of Stonehenge and its use through time. The combination of techniques employed has created a workflow which can be applied to a multitude of geographically and chronologically diverse heritage projects. The success of this project stands as a testament to the benefits digital technology can have on archaeological projects and has the potential to have a wide ranging impact on the way archaeologists perceive and utilise technology on future heritage projects. Marcus Abbott BA MIfA
Bronze Age rock art plotted on stones 4 (right) and 5 (left), it is now clear that artwork is placed in deliberate panels on specific stones
Luminance Lens derived image of stone 4, rock art features emit light
Super high resolution mesh created with Agisoft Photoscan
Plane Shading in Pointools, rock art features visible on stone 4
RTI the interplay of light and shadow defines rock art features on stone 4
Mesh of stone 4 in the RTI dome
Area of Neolithic stone working previously interpreted as a carving
Broken stones can be reconstructed in the virtual environment
Transverse tool marks on stone 59a note the ‘stop line’ at the right end of this image, indicating a deliberate boundary created by the Neolithic builders to work up to
Image of the laser scan survey point cloud data, collected by Greenhatch Group for English Heritage
The Beginning
The first results
Known prehistoric carvings
A detailed look at the 0.5mm mesh data
Visualising point clouds
A breakthrough in Photogrammetry
A new technique Luminance Lensing
Conclusion
Introduction
I am grateful for the continued support of Agisoft (Photoscan) Bentley Systems (Pointools) who have contributed software sponsorship for this and future projects. English Heritage for taking the innovative step to fund this research. Mike Pitts for publishing the exclusive article in British Archaeology. Cultural Heritage Imaging for open source RTI softwarehttp://www.agisoft.ru/http://www.pointools.com/http://www.english-heritage.org.uk/http://mikepitts.wordpress.com/http://culturalheritageimaging.org/shapeimage_45_link_0shapeimage_45_link_1shapeimage_45_link_2shapeimage_45_link_3shapeimage_45_link_4
Acknowledgements
Onsite examination was conducted to verify our findings, this stone, number 4 has over 50 Bronze Age carvings most of which are invisible to the naked eye.