Preserving endangered heritage across the world.
Heritage is about more than monuments. It is also about people: how they interacted with the buildings in daily life and how their sense of belonging has shaped them. This is why organisations such as UNESCO were established to protect the world’s cultural heritage from damage through natural disaster, neglect, vandalism, and human conflict.
One way to preserve monuments is to record them digitally, using specialist laser scans, to create 3D models. This is what the Curious Travellers Project does, using a resource that exists in abundance: the curiosity of millions of tourists and locals worldwide, who take photographs of what they see in the world around them.
We then augment these crowdsourced images – from digital cameras or scans of photographs, slides, and negatives – with data that is available free online and from social media sites to create 3D models of ancient monuments and sites, producing accurate representations without artificial or artistic reconstructions. Importantly, the project is more than just the 3D content. By using geospatial and archaeological data that describes the site within its landscape, its context is included, providing a lasting legacy that contributes to local historical environment records.
However, to be effective, we need to influence the way travellers record their holiday snaps: it takes more images of a site than you would think to create a 3D model. Crucially, we require views of every angle to show all perspectives – from the whole building to individual architectural features, as well as the back or unadorned sides of a monument, which people often do not think worth recording but are essential to provide a complete picture. Then, using special software, we can identify images that overlap, and the relative position of the camera is then calculated to transform 2D images into a 3D model.
So far, much of our work has concentrated on a few highprofile sites, including the 14th-century basilica of St Benedict in Norcia, Italy, destroyed by an earthquake last year, and Palmyra in Syria, attacked by Daesh. In these cases, media coverage generated a huge number of images to which we add from online sources. As more images come to light, we will incorporate the new data to further refine our 3D models.
Our conservation ethic means that we do not fill in gaps in the models with approximations of missing data. So, while our photo-realistic 3D models may not appear as complete as computer-generated (CGI) reconstructions, they are totally accurate representations of structures from which precise measurements or relationships can be retrieved. By working faithfully with these images, archaeologists can document weathering and erosion, changes in land-use, and gradual or sudden destruction over time – thus helping to inform longterm management of such sites.
But to succeed, we need your help. We would be grateful for your photographs of places you have visited, which you can upload via the project website (visualisingheritage.org). Particularly urgent are images of Palmyra (Syria), Cyrene (Libya), Amatrice and Norcia (Italy), Bagan (Myanmar), and Drummonds Mill (Bradford, UK). The project is also carrying out detailed work on Diocletian’s Palace (Split, Croatia) and Fountains Abbey (Yorkshire, UK). We can never have too many pictures, and it could be that yours are just the ones we need to fill the gaps and make the model work.
To explore the models, click here: www.visualisingheritage.org/cwa.php
Sources and further information:
Edward Faber, Tom Sparrow, Andrew Murgatroyd, Andrew S Wilson (Project PI), Vince Gaffney (Project Co-I), and Chris Gaffney (Project Co-I), , Bradford Visualisation, University of Bradford; Eugene Ch’ng (Project Co-I), University of Nottingham; Richard Bates (Project Co-I), University of St Andrews; Richard Cuttler, Director of MOSPA Heritage; and Gareth Sears, University of Birmingham.
Follow the project and see the developing 3D models on www.visualisingheritage.org, or find us on Twitter and Facebook: @curioustravell2