Technology in the desert: how to map in 3D Egyptian and Roman excavations using photogrammetry

Photogrammetry makes it possible to reconstruct the three-dimensional shape of an object or environment from photographs taken from different viewpoints. The principle is similar to human vision: by observing the same object from different positions, we can perceive its depth. Developed in analogue form in the early twentieth century, with the development of computers it first became analytical and then completely digital. Today we can process thousands of photographs and obtain highly detailed and metrically reliable three-dimensional models.

On missions in Egypt, we use photogrammetry mainly at ground level and in underground or very confined environments, such as the Tomb of Kha and Merit. We also work in difficult conditions: wind, dust, high temperatures and very intense light can affect the equipment and the quality of the data. For this reason, we use special lenses and follow very rigorous acquisition geometries, which are essential for obtaining metrically reliable three-dimensional models. For surveying the burial shafts leading to the burial chambers, I built a metal cage to hold the camera, equipped with a flash and a fisheye lens that allows us to photograph a large part of the shaft walls even at close range. The metal structure is lowered, parallel to the walls and also along the four corners, into the shafts by using ropes and photographs are taken remotely using a remote control. We mark the ropes at regular intervals, in order to make sure that the acquisition spacing doesn’t vary excessively between the different descents; in fact, the operation has to be repeated at least eight times for each shaft, reaching as far as 20–25 metres underground!

All the images are imported into a photogrammetry software and, through preprocessing, we make sure that all the data collected are sufficient; otherwise, we supplement them with additional photographs. At the same time, on the surface, we use a precision optical instrument, the total station, to take a small number of essential coordinate measurements, and it must remain absolutely stationary. Through its telescope, we measure the centre of several coded markers that we place within the area we’re surveying, and that are also “recorded” in the acquired images. When entered into the photogrammetry software, these measurements provide us with the accuracy of the final three-dimensional model, which for us is acceptable only within the range of one centimetre. The same operation can also be repeated underground, provided there is sufficient space in the subterranean burial chambers. It’s different for shafts that are vertical, narrow and extend for several dozen metres: in this case, we adopted a solution as simple as it is effective, we use bottle caps with a nail in the centre, so that they remain firmly fixed in the rock walls. The coordinates of these “improvised” markers are calculated and checked afterwards and help us understand how far we have reached, so that the following year we can start again from where we left off.

Around a hundred photographs may be enough for a room. For an above-ground temple-tomb, such as that of Horemheb at Saqqara, we may reach a thousand. To document the entire Roman site of Umm al-Dabadib, we took more than 10,000 photographs.

We use a software that’s also very commonly used in architecture and engineering courses, and we make sure that we carry out rigorous numerical checks before building the final model. For this reason, in addition to photogrammetry, we use topographic surveying. Every year we create three-dimensional models of the excavations, linking the new surveys to the topographic network, so that all the models share the same reference system and can be perfectly superimposed. If I open them in the same viewer, I can see how the site has changed over time. This is the added value we offer to Museo Egizio: a digital diary documenting the entire evolution of the excavation, allowing us to go back in time through the different stages. It starts with the site covered in sand, then the first columns emerging, the shafts, the burial chambers, all the way to the complete tomb. As our colleagues at Museo Egizio always remind us, archaeology is unfortunately a destructive science: at Saqqara, for example, the first levels of the excavation corresponded to Coptic culture, but to reach the Pharaonic period it was obviously necessary to remove those layers. They no longer exist physically, but they survive digitally thanks to our three-dimensional models, which allow archaeologists to take measurements, study and verify, preserving the past.

Yes, in the field you also have to know how to improvise. During one of the first missions, when we were still camping in the desert, the charging cables for the camera batteries broke: we had some solder, a soldering iron and, in the darkness of the night, with only our head torches, we connected it to the off-road vehicle’s batteries, soldered the cables and managed to finish the job! It is one of the aspects I love about my work: finding solutions in difficult contexts.

The choice of photogrammetry also stems partly from practical needs. I also specialize in the use of laser scanners, but taking this type of equipment to Egypt can be complicated. In 2014, for example, some instruments were held at customs because they were considered “non-standard”. Another possibility would be to use drones. I obtained my pilot’s licence in 2015 and used them during my PhD for environmental surveys, to reconstruct landslide scarps and monitor cultural heritage. In Egypt, however, the use of drones is prohibited for security reasons. So we found an alternative solution: we pretend to be drones ourselves! We use a photographic pole more than four metres high, with a professional camera mounted on top. We have to take more photographs than would be necessary with a drone, but we obtain images with very high resolution and accuracy. This is particularly important at the end of the mission, when we carefully document the situation reached, before covering the excavation back up. Those images will provide the starting point for the following year.

At the end of an excavation, an operation known as backfilling is carried out. To preserve the site, the shafts are closed, sheets are placed over the layer that has been reached and everything is covered with sand. The following year, work resumes more or less where it left off: this is why the missions are so long, lasting from four to six weeks. There are a great many activities and they often cannot proceed in parallel. The final week is the most important: before closing the site, we have to make sure that all the documentation for the digital diary is complete.

During the missions we also work alongside other experts: at Umm al-Dabadib, in addition to me, Professor Corinna Rossi and Clementina Caputo, we are joined by two archaeologists, Stefania Alfarano and Cesare Iezzi, and Nicoletta De Troia, an expert in social and cultural relations in Late Antique and Early Medieval Egypt. 

At Saqqara, for example, the team consists of around twenty experts, including ceramic experts, illustrators, Egyptologists, archaeologists and anthropologists. Every single find is recorded and catalogued under the supervision of the Egyptian Antiquities Inspector, who accompanies us on every mission. Everything is then transported to a storeroom. At this stage, I often work alongside the excavation illustrators, who mainly draw ceramic objects or worked stone fragments, including hieroglyphic inscriptions. To support them, I carry out three-dimensional surveys of the inscribed surfaces, and from the three-dimensional model I can produce an ultra-high-resolution orthophoto, which can subsequently be traced using photo-editing software. The added value is the objectivity of the data, which are not affected by the illustrator’s interpretation or personal technique.

A normal photograph is, by definition, a perspective representation; depending on the lens used, there’s always a certain degree of distortion, so it cannot provide metric information. An orthophoto, on the other hand, is an orthogonal projection generated from a three-dimensional model. In this way, any part of the image can be measured and is free from distortion both at the centre and at the edges. When it comes to objects, having a 3D model of, for example, a jar clearly allows me to automatically extract profiles and sections at any position, greatly speeding up the work of illustrators, who instead produce a limited number of drawings, which may in any case potentially be influenced by subjective interpretations.

Yes, and it could become increasingly so. One initial area of application concerns the organization of archaeological databases. Over time, different cataloguing systems have been used, and the same type of find may be recorded under different names. Artificial intelligence can help us identify matches, eliminate overlaps and connect information from different archives. Looking ahead, if we have a large number of three-dimensional models of similar objects, we could also use these tools to rapidly perform statistical analyses and identify relationships between finds discovered in different areas of Egypt, for example by recognizing objects that may have come from the same production area.

This year, at Umm al-Dabadib, inside a hole in the floor of a house, we found around twenty glass “unguentaria”. None had ever been found before in this part of the desert. I spent entire days putting the fragments back together!

But one of the most curious discoveries was a sort of “little treasure” hidden inside a ceramic container. These vessels were used to hide and bury possessions considered valuable, with the intention of retrieving them later. The site was probably abandoned when the underground aqueducts that made it possible to cultivate vast areas of the desert stopped supplying water, and no one ever returned to retrieve that container. Inside we found bracelets, ivory beads, coins bearing the effigies of various emperors and four ivory gaming dice. When I took the dice to the storeroom to photograph them, I began turning them around to document all six sides. That was when I noticed that on one of the dice the number four appeared twice, while the number three was completely missing. It was a loaded die! It was a very amusing discovery for everyone.

Actually, I went… from marble to sand! During and after my PhD years, I spent seven years in Milan Cathedral. Thanks to the dedication and commitment of Professor Cristiana Achille and Professor Francesco Fassi, I had the opportunity to develop professionally and scientifically among the spires of the Duomo. I worked on photogrammetric and laser-scanner surveys and on three-dimensional modelling of the Main Spire and the interior of the cathedral, in support of the restoration work. Being from Milan, it was a great source of pride for me! Over time, however, I increasingly wanted a change, to take on new challenges and bring these technologies into completely different contexts. And so, from Milan Cathedral, I arrived in the Egyptian desert, venturing as far as the borders of the Roman Empire at its greatest extent.

We will soon be leaving for Egypt again. From 2 to 30 October, we will be back in Saqqara for a new archaeological mission, during which we’ll continue the work begun in previous years, and document the new phases of the excavation. As every year, we know where we will start from, but we cannot know what we will find beneath the next layer of sand…