The 3D survey at El Fuerte de Samaipata was conducted over the course of three days in July of 2006. An Optech ILRIS 3D scanner was used to measure and record tens of millions of survey points in a semi- automated fashion across the surface of El Fuerte. The scanner collects data points for surfaces that are within its range and field of view and therefore the scanner must be repositioned at various points around its target.

The survey of El Fuerte required the collection of over 70 scans from 25 positions around the rock outcrop to ensure that data was collected over as much of its area as possible (see above). These scans recorded survey points across the surface of the rock outcrop(below left) with each survey point recording a point in 3D space (i.e., recording X, Y, and Z coordinates). Thus, each of these scans yield what is known as a "point cloud" (below middle), a dense array of survey points, often numbering in the tens or even hundreds of thousands for each scan, that effectively characterize the surface of El Fuerte as a three-dimensional computer model (below right) and provides an accurate representation of the site at the time it was scanned. In total, over 85.5 million survey points were collected at El Fuerte de Samaipata, covering an area measuring 90 m x 200 m x 40 m.

Scanning El Fuerte presented several challenges that held ramifications for both field survey and data processing. For instance, operation of the Optech scanner is best from a high, distant vantage point. This ensures that large areas can be scanned with the Optech scanner as close to perpendicular to the surface as possible and free from obstructions such as vegetation. Given the topography of the area, this was seldom possible, as El Fuerte IS the high vantage point. While some scans were acquired from the park's miradors (viewing platforms), the majority of the scans were acquired from the ground surface around the monument as well as from on the monument itself. Both of these scanning positions introduced their own unique challenges.

Scanning from the park's miradors provided the only opportunity for the survey team to scan El Fuerte from above. While scanning from the miradors provided the necessary distance from the monument to acquire scans of large areas, there were not a sufficent number of positions available to collect data across the entire site. With only two miradors and two elevated boardwalks, all on the south and west sides of the monument, data from the north and east sides of El Fuerte are not as complete. Another challenge related to scanning from these platforms, although not as prevalent, was their slight movement due to the strong wind gusts that sweep the mountaintop. This movement of the scanner during data collection contributed to slight errors in the scan data.

Scanning directly on top of El Fuerte allowed for areas to be surveyed that were otherwise omitted in the scans acquired from other scanning positions around the monument. Special care was taken to ensure that damage was not done to the fragile surface either by the survey team or the survey equipment. All team members removed their shoes during data collection on the monument and pads were used to shield the stone surface from the pointed metal legs of the scanner tripod. While scanning from the surface of El Fuerte provided access to areas that would have been otherwise unsurveyed, it often required that the scanner was positioned at an oblique angle to the monument's surface. This oblique angle of survey results in the loss of some data, as the laser beam that measures the survey points does not always return to the scanner to be recorded, but rather is reflected away from the scanner.

Each of the individual scans collected during field survey have contributed data to a unified model of El Fuerte de Samaipata. These scans have been parsed, cleaned, segmented, assembled, and meshed to create a model that represents approximately 70 percent of the surface of the monument. The parsing process converts the data from the proprietary format of the Optech scanner during data collection. Cleaning removes extraneous data, such as that from vegetation, areas scanned beyond the monument, etc. The scans of large areas are then segmented into smaller sub-scans for the purpose of importing into the processing software. Processing was conducted using InnovMetric's Polyworks software suite. This software utilizes advanced algorithms for all stages of processing 3D scan data. Segmenting the large scans enables the operator to tailor each algorithm's parameters to better fit the parameters of the data. Most importantly, scan segmentation allows for the compensation of beam divergence, the property of the laser beam that is measuring the 3D survey points to spread as it travels across large distances. By segmenting the scans, a better quality surface can be derived from the 3D survey points. Below is an example of segmentation applied to a scan. The image on the left is a complete scan over the site. The image on the right is the same scan, segmented by range (distance from scanner), to allow a more accurate surface to be computed.

Assembling the scans entails manually positioning scans to the same relative orientation (below, left) and matching common points between each of the individual scans (below, right).

These points bring the two scans into relative alignment (above, left) based on the common points chosen. This is followed by the aplication of complex computer algorithms that consider all points in the scan and 'fits' the scans together very precisely (above right). While the shift made between the scans is minute in the area of overlap, consider the effect of this shift on survey scans meters or tens of meters away.

After the individual scans were assembled and precisely aligned, the process of overlap reduction was undertaken. Overlap reduction serves to eliminate redundant data between scans. While overlapping areas are necessary to match the individual scans together as demostrated above, this data can be removed once the scans are assembled. Care was taken during the overlap reduction process to retain the most accurate data possible.

Excerpts of the processed data is available for download on the Survey Results section of this website.

All original survey data has been archived and can be made available to researchers and other interested parties.