Preserving Ancient Egypt with Laser Scanning
Aerial view of the King's Palace mounds of the Birket Habu, an artificial lake constructed for use with the palace of Malqata. Photo courtesy of Joel Paulson.
Under the auspices of the Joint Expedition to Malqata (JEM) and with the permission and assistance of the Egyptian Ministry of Antiquities (MSA), the survey team, under Joel Paulson of NV5 (San Diego, CA), performed a laser scan to create a detailed map of the Kings Palace at Malqata, one of the few palace sites known from Ancient Egypt.
The West Bank of Thebes, modern Luxor, in Egypt is one of the richest archaeological regions in the world. Boasting famous and important sites such as the Valley of the Kings, the Valley of the Queens, the tombs of the nobles, as well as important temples at Deir el-Bahari and Medinet Habu, the region is justifiably popular with tourists. But the West Bank also includes many lesser-known, but still highly significant archaeological treasures. One of these lesser known and certainly less visited sites is the palace at Malqata, at the south end of the West Bank archaeological zone.
This locale is actually an entire complex of palaces, estates of the nobility, villages for laborers and staff, a temple, an artificial lake called the Birket Habu, and some poorly understood outlying structures and features. But surprisingly for the size and complexity of this palace, it was built and apparently used by only one pharaoh, Amenhotep III, who reigned from ca. 1350-1325 B.C.
Since 2009, the Joint Expedition to Malqata (JEM), consisting of a team from the Metropolitan Museum of Art (New York), the William Carlos Museum at Emory University (Atlanta), the Ancient Egyptian Archaeology and Heritage Fund and independent scholars, has been working at the site to study, catalogue and conserve the various components of the site.
The site of Malqata is massive, encompassing approximately 1,300 hectares (3,200 acres). At the north end is the temple of Amun, with its large courtyard and numerous revetments, used as foundations for the inner chambers. Just to the south of the temple is a pavilion with a raised mud-brick structure almost perfectly square, perhaps used as a seating platform for the pharaoh for rituals. To the southeast is the King’s Palace. This structure went through at least three building phases and contained large pillared halls, as well as what are apparently throne daises and private chambers, including bath facilities. Just west of the King’s palace are a set of large houses, probably for use by the officials of the palace or other upper-class people during their stay at the palace for royal functions.
East of all of these is the massive artificial lake known as the Birket Habu. In ancient times, the pharaoh had this lake created solely for use of the palace for its ceremonial functions. The lake measures almost three kilometers long from north to south, and two kilometers wide from east to west. It was fed by a canal from the Nile River on its east side. This lake was so massive that modern Egyptians have used the back piles of dirt that were excavated from it to raise their villages above the flood level of the Nile. This construction has obscured the Birket on several sides in satellite and aerial images, but it can still be traced by the villages themselves. The west side of the Birket still has the original mounds intact, although modern expansion threatens them.
Many other structures, including secondary palaces, villas of the wealthy and supporting villages, lie around the King’s Palace, and the site of Malqata continues southerly from these structures for another five kilometers, although the features grow sparse as the site continues south. The palace and its associated buildings and features create an expansive ritual landscape for ceremonies that were performed for the pharaoh to renew him as he grew older and less active.
Although many tombs and temples have been excavated, examined and studied throughout Egypt, archaeologists have identified and excavated very few palaces from ancient Egypt. This is especially surprising given the importance of the institution of kingship to the ancient Egyptians, making each of the known palace sites highly significant for their insight into that part of the culture.
The site of Malqata has been known since the beginnings of Egyptology. In the late 18th century, French scientists who traveled with Napoleon’s expedition included the Birket Habu and some remains of the palace on their map of the west bank. About 100 years later, in 1898, the Egyptian Antiquities Service conducted the first known excavation at the site of the palace. More excavations followed.
An American, Rob Tytus led excavations there in 1901 and 1902. Shortly thereafter, between 1912 and 1919, the Metropolitan Museum of Art took over the site, excavating the palaces, villages and the temple, with a hiatus in 1915 and 1916 because of World War I. From these excavations, the archaeologists discovered a number of colorful pieces of decoration in the King’s Palace, some of which are now at the Metropolitan Museum of Art in New York and in the Cairo Museum. But amazingly, some of the original decoration still exists at the site, though mostly in small fragments from its collapse.
Yet, for all this previous work, more remains to be done at Malqata. Except for some door jambs, thresholds and a few other features, almost all of the construction of buildings at Malqata was mud-brick. Since the Metropolitan Museum excavations of the early 20th century, the mud bricks of the structures have been subject to weathering, and, because of climate change in Egypt over the last 50 years, the erosion of the bricks has quickened. One of the primary goals of the Joint Expedition to Malqata is to check the architectural plans drawn by the original archaeologists and to bring them up to modern standards. Following this step, the original walls are being capped with modern mud bricks to conserve the ancient walls and prevent further erosion. It was to assist in this portion of the work that we thought it would be vital to perform a 3D laser scan of the existing walls to record them prior to capping them.
In 2012, the Metropolitan Museum of Art hired my employer, NV5 (San Diego), to perform the laser scan. Laser scanning as a technology has been around for surveying since the late 1990s, but, because of the costs of the equipment and the specialized software and skills required for managing the millions of points obtained from the scanner, this technology has not been used as extensively for archaeology as it could be. But archaeology is a perfect application for the scanners.
At its simplest, a high density survey (HDS) scanner is just another piece of surveying equipment; it measures points based on angle and distance from the centerpoint of the scanner, so each point has a known XYZ coordinate relative to the scanner position. The HDS scanner just does it very quickly. Some of the more modern scanners can collect over 1,000,000 points per second. The collection of so many points creates a dense set of points – the pointcloud – that represents the subject at a true scale in three dimensions.
With software, the user can combine scans from multiple locations, linking them together by targets common to the different setup stations. By setting up on known points and using a known backsight, or by resection from several known points, the scan data can be linked to an overall coordinate system and the results compared to total station and/or GPS location data. Depending on the density of the points and the size of the subject, a laser scanning project can easily consist of millions of points, and, sometimes, for more complex scans, billions of points.
For the Malqata project, I used a Leica C10 laser scanner, capable of collecting up to 50,000 points per second. Because of the plans to cap the existing architecture at the King’s Palace, we concentrated on that locale for the scanning. Around the palace, I set a series of survey points using a total station to act as control for the laser scanning. To include the scan data into a GIS with real world coordinates, I based the control in terms of UTM, Zone 36 based on GPS surveys I had performed in 2008 and 2010.
From each scan station, I included at least three targets of the control for initial alignment of the scans. The palace contains many rooms and a number of setups were needed to ensure that I included the mud brick walls all the way to the ground in each one of them. In the end, for the 21,330 square meters of the palace that we scanned, we had 55 setups.
The entire sequence of scanning took five days, and for the king’s palace we ended up with approximately 548,535,715 points (more or less) in our pointcloud.
As those familiar with laser scanning know, the real work is in the processing of the scans, not in the fieldwork. After we had gathered all those points, we had to do something with them to make it all worthwhile. After getting back to the U.S., I registered all the scans together using Leica’s Cyclone software to create a single model from the points.
For the registration, I started with the alignment provided by the scans of the targets on the known control points. I would add additional control points to the scan from points that were identifiable from one scan to another. Then I would allow the software to run its “scan to scan” automated registration. At each step of the way, I checked the RMS (Root Mean Square) values obtained from the registration. If the values were higher than the allowable minimum for the project, I would perform a combination of actions in the software to add more control points, refine the previous control, or disable the control points until the RMS values were within tolerance.
In addition to checking the RMS values, I performed a visual check of the results by putting the points from each scan into a different color to show which station it had come from. Then, in the software, checking at applicable walls and floors, the distances from a sample of points from one scan to a plane created by points from a different scan. This confirmed that the points from each scan were within tolerances relative to points from the other scans.
Then I started cleaning the pointcloud. Cleaning the “noise” from a pointcloud is never as quick and easy as it seems like it should be, but a busy archaeological site makes it an even greater challenge. As the scanner was collecting data, workmen were going back and forth carrying away the sand and dirt, re-clearing the palace walls; archaeologists were sitting on the walls drawing the mud brick details; photographers were working around the sporadic finds of more painted plaster. All of these people had to be carefully “cleaned” from the pointcloud, along with their equipment, so that we ended up with only the walls and the ground.
From the scan and the on-going study of the King’s Palace, the JEM team has made a number of refinements to the original architectural plans. We have also proceeded with the work of capping the ancient mud brick walls with modern bricks, knowing that we have a complete and accurate model of what exists beneath the cap.
Another goal of the JEM project is to create explanatory signage and images at the site to attract tourists, without them causing further deterioration of the site.