Data Set Specifics

Movie of Machu Picchu, Peru data:

1.7 mb Flash 8.0 format

The movie requires the Flash player to view, download the free player here.

A photo journal of the Tiwanaku portion of the project is availalble at: http://www.cast.uark.edu/invirmet/photo_journal_19.htm

 

Summary of Laser Scanning Practices at Tiwanaku and Machu Picchu

In the summer of 2005, researchers from CAST traveled to South America to scan the world heritage sites of Tiwanaku , Bolivia and Machu Picchu, Peru. Settled high in the Bolivian mountains along a flat uprising of land referred to as the altiplano, the city of Tiwanaku was settled around from A.D. 500 – 950. The Tiwanakan people constructed large mounds, platforms, and stone structures to form the core of their city whose influence stretched across the altiplano and into the Andes. Settled before the time of the Inca, Tiwanaku was recognized as a birthplace for the people of the Andes when the site was rediscovered by Incan conquests in the 1500's.

Map taken from http://www.archaeology.org/interactive/tiwanaku/map.html
(Click image to view larger size)

With the emergence of the Inca and nearly 350 miles away, the city of Machu Picchu was built high in the Andes Mountains outside the modern day city of Cuzco. Constructed in the late 1400's by the Incan ruler, Pachacuti, Machu Picchu existed as a place of worship and as a royal retreat for Pachacuti and his guests. Abandoned only decades after its construction, Machu Picchu was lost to the jungle only to be rediscovered by Hiram Bingham in 1911.

Given the cultural significance of both Tiwanaku and Machu Picchu, both sites were considered excellent candidates for laser scanning.

Scanning at Tiwanaku

The “monumental core” of Tiwanaku contains several large stone monuments. These include the Akapana Pyramid, the Putuni, the Kalasasaya, the Semi-subterranean temple and other smaller structures and monoliths. During the 2005 field season the Optech ILRIS-3D laser scanner was used to record much of the core area of Tiwanaku. Scans of the Putuni, the Kalasasaya, and the Semi-subterranean temple were obtained at a 3 cm resolution. Other areas of interest, including a GPR survey area, an excavation trench, and several stone monoliths, were also scanned at higher resolutions of up to 2.5 mm.

The Akapana Pyramid was not scanned largely due to the size and height of the monument and also due to the vast amount of vegetation covering the surface. Vegetation is problematic in laser scanning as it can completely impede the laser from obtaining measurements of a feature of interest. It is recommended that a different technique such as that employed by Markussen et al. be used for mapping the Akapana in future fieldwork at Tiwanaku.

Due to the data acquisition range of the Optech, a high, distant vantage point is often ideal for scanning large structures. This allows one to acquire a lot of data in relatively few scans. Scanning from a distance can also take considerably less time as it reduces the number of scan setups thereby significantly reducing the overall time in the field. At Tiwanaku, the primary vantage point was from atop the Akapana pyramid. The view from the Akapana encompasses the majority of the Tiwanakan landscape and it provides a near birds-eye view of the monument area.

Panoramic view from the Akapana Pyramid showing the Kalasasaya (left) and Semi-subterranean stone structures (right)
Image courtesy Snow L. Winters (Click image to view larger size)

At the beginning of the project four scans were acquired from atop the Akapana from a single scanner position. The scans were acquired at a 3 cm resolution taking an average of 1.5 hours each and containing approximately three million data points each. The four scans aligned provide a nice perspective for viewing the relative positions of the Putuni, the Kalasasaya, and the Semi-subterranean temple.

View of scan data from Akapana Pyramid depicting the Putuni, Kalasasaya, and Semi-subterranean structures at Tiwanaku
(Click image to view larger size)

After viewing the results of the first few scans, it was decided that a 3 cm resolution was appropriate to meet project needs. It provided adequate detail of the design and architecture of the structures and the scans were also acquired in a reasonable time frame. There is always a critical balance between scan resolution and scan acquisition time that the scan operator must evaluate. In reality time constrains the volume and resolution of data that researchers are able to collect. Therefore, resolution must carefully be evaluated taking into consideration the level of object detail that needs to be discerned in a scan and the amount of time that can be spent in the field.

The Akapana Pyramid provided the only high vantage point for scanning at Tiwanaku. While the prospect of mounting the scanner on scaffolding was considered, given the lack of time to test such methods first, it was decided to obtain the rest of the scans from the ground. This was done by positioning the scanner at numerous positions around and inside each of the monuments. The largest of the structures at the site, the Kalasasaya was the first structure to be mapped in its entirety.

A large portion of the Kalasasaya in particular the southern walls and the interior floor was obtained in the scans taken from atop the Akapana. The remainder of the east, north and west walls and of the interior of the structure were scanned in a series of the 105 scans. Because the scanner was positioned much closer to the structure these scans took considerably less time to acquire. The average scan time for scanning the Kalasasaya at ground-level, close range and at a 3 cm resolution was 60 seconds. The remainder of the scans of the Kalasasaya were acquired in four days collecting an average of 30 scans a day.

One issue that came up during the scanning of the Kalasasaya was the parallax issue between the laser scanner and the internal camera. The field of view (FOV) of the laser and of the camera converges at a range of 150 meters. In other words, the laser and the camera at ranges less than 150 meters do not “see” the same view of a given scene. As a result, there is often a slight shift in the laser scanner data when compared to the image taken by the scanner of the scan area. This parallax issue becomes more apparent when the distance between the scanner and the object being scanned becomes shorter.

Notice the areas in red on the outer edges of the image (left) are not present in the scan data (right)
(Click image to view larger size)

On the second day of scanning the Kalasasaya from the ground, scans were acquired with only a minimal degree of overlap deemed sufficient for the alignment procedure. However, after data processing it was apparent that although the images showed sufficient overlap there was no real overlap in the data itself. This prevented image alignment so new scans had to be obtained and virtually a day's worth of work had been lost. To remedy this issue, all of the remaining scans were taken with a significant degree (30-40%) of overlap.

Scans of the Putuni and of the Semi-subterranean temple were also acquired. The Putuni was somewhat difficult to scan because it is a stepped, sunken temple that includes interior and exterior walls as well as two ground surfaces. Again two high points were used, one from the Kalasasaya and the other from a mound to the north of the Putuni to scan as much of the Putuni as possible. Overall, a total of 31 scans were acquired of the Putuni at a 3 cm resolution. The Semi-subterranean temple was relatively easy to scan requiring only two scans for each wall of the temple. Eight scans were obtained of the Semi-subterranean temple at a 1 cm resolution.

Several of the stone monoliths including the Puerte del Sol (“Gateway of the Sun” ) and the Ponce monolith both found in the interior of the Kalasasaya were also scanned at a finer resolution of 5 mm. A 5 mm scan resolution was chosen in order to attempt to discern the high level of artistic detail in the design of each structure. Other areas of interest were also scanned during the last few days at the site.

Scan summary from Tiwanaku

Overall, a total over 200 scans were obtained at Tiwanaku resulting in a collection nearly 75 million data points. Data collection took place over a period of eight days. Scans were obtained of the Kalasasaya, the Putuni, and the Semi-subterranean structures as well as of two stone monoliths including the Puerte del sol and the Ponce monolith. Scans were also obtained of an excavation trench located to the west of the Akapana Pyramid and of the 2002 GPR survey area also located to the west of the Akapana.

Scanning at Machu Picchu

Machu Picchu was considerably different to scan and as a result a number of issues were brought up that had not been addressed in previous scanning projects. The layout of Machu Picchu is unlike other any site in the world. It is located on the top of a steep mountain that has been cutout to form a series stone terraces and structures. The structures within the site are located so close together that it is impossible to put a long-range scanner such as the Optech between the structures to obtain scans of the entirety of each and every wall. An undertaking of this proportion would consequently take months rather than days. As a result, most of the scans of Machu Picchu were acquired from several distinct vantage points that provided good panoramic views of the site. However, because there were a limited number of scan locations within the site there were consequently a limited number of scans that could be acquired using the given scan unit. Therefore, the scans of Machu Picchu do not necessarily exhibit the detail of every wall of every structure but instead they show the general layout of the entire site at a relatively moderate level of detail (3 cm).

Image of Machu Picchu illustrating the proximity of the stone structures
(Click image to view larger size)

The Optech laser scanner operates in the infrared range making it possible to scan at any time of the day or night. It was necessary to scan Machu Picchu after the park closed because there were too many tourists that would interfere with the scans during the day. Scanning typically occurred between 4:30 pm and 10 pm each evening and each scan took about an hour. The first day at the site was spent touring the site and noting ideal scan positions. An estimated 3-4 scans were acquired each evening, although this number increased as new methodologies for scanning at night were developed.

The Machu Picchu project was the first of its kind where laser scanning took place in the dark. Even though laser scanner setup locations were identified during the day, once the scanner was setup in that location at night it was impossible to determine its field of view (FOV) until partway through the scanning process. If the FOV was incorrect then the scan was stopped to adjust the angle of the scanner before completing the scan. This was obviously not the most time-effective solution.

Instead it was determined that conducting a quick, course scan provided a nice outline of the features in the scanner's FOV. The scan resolution was set high (often greater than 100 cm) making the average scan acquisition time less than a minute. In viewing the pre-scan, the scanner angle and position could appropriately be adjusted. This technique was also useful for determining the correct amount of rotation between two successive scans. Again without being able to “see” the scanner's FOV it was difficult to determine how far to rotate the scanner between scans for a sufficient overlap in the data. Initially, a light source was used to illuminate a common feature of interest between two successive scans. However, this method was more exhaustive and it was determined the same could be accomplished by using the above stated method.

Map of the six principal scanner locations, map courtesy Enjoy Peru Travel
(Click image to view larger size)

The scanner was setup at six principal locations within the site of Machu Picchu . In referencing the scan map below, positions 1, 3 and 5 were used to acquire generalized scans of the entire site.

Scans from positions 1 and 3 were acquired at a 3cm resolution while the scans from position 5 were acquired at a much courser resolution of 20 cm due to distance and time constraints. Scan 2 provided an additional perspective for viewing the military sector. Positions 4 and 6 were setup to scan the sacred hill and the temple structures respectively. These features were of particular interest to archaeologists and site conservationists that were part of the team.

In addition to scanning the ruins of Machu Picchu , scans were also acquired of the ruins on top of Huayna Picchu, the steep mountain to the north of Machu Picchu . The image to the right depicts the scanning of Huayna Picchu which was an estimated 500 meters away from scan location 5. The 3-Dimensional recording of this otherwise difficult to map structure was made possible given the long acquisition range of the Optech scan unit. Scans of Huayna Picchu were acquired at a 3 cm resolution and took approximately 2 hours to obtain.

Scan Summary from Machu Picchu

A total of four nights were spent scanning at Machu Picchu. Twenty five scans were obtained resulting in a collection of over eighty million data points. Without the use of the camera CCD screen, the team overcame a number obstacles dealing with the issues of laser scanning at night. Once these issues were resolved however, scanning continued as normal and a lot of data was acquired in four days time.

Topographic map of Machu Picchu, map courtesy Enjoy Peru Travel
(Click image to view larger size)

The work reported here has been supported, in part, by NSF via Grant BCS-0321286 Acquisition of a High Accuracy/Resolution Landscape and Structure Characterization System (HARLS-CS) for Anthropology, Archaeology,
Architecture, Biology and Geosciences and Grant IIS-0431070 Computing and Retrieving 3D Archaeological Structures from Subsurface Surveying

 

......
Home | View the Data | Instrumentation | Projects | How we do it | Collaborators

Center for Advanced Spatial Technologies, University of Arkansas
Phone: (479)575-6159 | Fax: (479)575-5218 | Email: invirmet@cast.uark.edu
......