Terrain LIDAR scans jungle in search of a lost city
I am presently reading a very interesting book entitled, The Lost City of the Monkey God. This book is about a real-life archaeological adventure in which one of my favorite authors, Douglas Preston, is involved.
This jungle expedition has a technical aspect that got my attention---the use of airborne light detection and ranging (LIDAR) on an incredibly dense and unexplored, 20,000 square mile patch of the largest untouched rainforest in Central America, in a region known as Mosquitia. This area spans across both the Honduras and Nicaragua borders.
It would normally take an entire career to hack through a jungle like this with a machete in order to map this area, and that’s if they even had an inkling of what archaeological find might be there. Only legends existed of a great city in the mountains of the broad, unexplored regions of Nicaragua and Honduras. Explorers were said to have been searching for this city since the 1500s. Even explorer Steve Elkins was on a quest for this city, also known as the ‘White City,’ since the early 1990s. Elkins launched his latest expedition in 2012, except this time he had the laser mapping capability of LIDAR.
Acquiring the LIDAR equipment
LIDAR was first used for mapping in about 2009 when a husband and wife team of archaeologists, Arlen and Diane Chase, got together with NASA and the National Center for Airborne Laser Mapping (NCALM) to map an area called Caracol in Belize. Previously only radar and satellite imaging was available to do such mapping but could not penetrate through a dense jungle canopy. The best ground resolution using radar mapping in the mid- to late-'90s was only around 90 feet. LIDAR could give a resolution of 3 feet, even under a jungle canopy. It would shoot thousands of laser bursts per second, creating a detailed topological map.
Elkins met with the staff of NCALM and managed to convince them that he could use LIDAR, which had never been used before as a tool in a pure archaeological expedition to look for something that no one even knew if it existed. The co-principal investigator and chief scientist at NCALM was William E. Carter, one of the fathers of LIDAR. Carter helped design and operate one of the first lunar laser ranging stations for NASA’s Apollo missions in the '60s. This system was able to measure the Earth-Moon distance to an accuracy of a few centimeters.
An old Cessna Skymaster plane, with its interior removed, would house the million-dollar LIDAR system that could peer right through the dense jungle canopy to identify and map the contours beneath. See a LIDAR design on an aerial drone here.
The LIDAR scan, able to map the ground even through a dense rain forest, ultimately revealed an area seemingly with rectangular-shaped structures as well as two perfectly linear lines and a right angle (Figure 1).
Figure 1 The LIDAR scan of the area in the Honduras jungle (Image courtesy of CBS News)
Now the expedition could begin since they know exactly which area to target a path.
LIDAR scanning technology1
The first step in the mapping and location of geodetic images, of the Mosquita region, was to install continuously-operating GPS base stations within 100 km of the remote areas needed to be mapped. This would enable getting high accuracy aircraft trajectories critical to get LIDAR data accuracy (Figure 2).
Figure 2 A project Map showing the three GPS basestations (Image courtesy of Reference 2)
Previously, space borne synthetic aperture radar (SAR) tried to penetrate this jungle canopy with only low-resolution pixilated images that did not provide enough useful detail to identify what was down there.
Next, a flight plan needed to be done that would enable adequate laser penetration of the jungle canopy so they could get maximum ground return signals. The flight plan would be structured so that it would maximize the number of reflected ground return laser signals through the thick jungle canopy to obtain usable images. So, the LIDAR sensor was configured and flight plans were created to scan every square meter of the rainforest from four different angles. This provided a multi-pass full service illumination.
Nominal flying altitudes of 600 meters above ground level (AGL) at a ground speed of 60 m/sec were achieved with the laser firing at 125 kHz with a divergence of 0.8 mrad (a 0.48 m footprint diameter). The team also tested higher laser pulse repetition frequencies (PRF) to obtain higher shot densities; however, the reduced energy in each pulse, in this instance, resulted in fewer detectable return signals from the ground. The scanning was done at +/− 15 degrees at 60 Hz and the final operating parameters achieved a minimum shot density of 25 shots/m2.
Over a two-week period, seven flights were made which totaled 32.1 hours with 8.4 hours of Laser-On-Time. A total of 3.5 billion laser pulses were fired over an area of 122.8 km2. Only 87 million actual ground returns were classified (most reflected off the dense canopy of vegetation). These returns were processed with an algorithm described in Reference 3. From the filtered ground returns, bare-earth digital elevation models (DEMs) were generated. From these DEMs multiple data products can be generated. The most common was a set of shaded relief images that represent the surface terrain stripped of any dense vegetation (Figures 3 and 4).
Figure 3 An image of a targeted river valley where human modification of the landscape is evident with a number of mounds arranged in a rectangular and other geometric pattern (Image courtesy of Reference 2)
Figure 4 A shaded relief image from a DEM of part of a river valley. Mound-looking figures can be recognized on either bank of the small river. These banks are separated by 600 m and in the flood plain, multiple lower relief archaeological images are identified (Image courtesy of Reference 2)
See the following video for more details:
A main disadvantage of LIDAR is the upfront costs to archaeologists which are typically higher than most research grants. The caveat is that if traditional field survey methods were used in this kind of environment, the quality of the data would not be as good and the time would be far greater to obtain only questionable results.
In 2014, archaeological excavation methodology was fundamentally changed when daily aerial and terrestrial structure-from-motion (SfM) was introduced2. This effort led to massive 3D datasets composed of point clouds, textured meshes, and gigapixel resolution orthophotos which were able to be rendered simultaneously and over different days of excavation.
Maybe someday soon we will find a way to coordinate unmanned aerial systems (UAS)/drones to perform these airborne tasks coupled with GPS coordinate and guidance systems. I would look forward to that. There are presently geospatial and surveying solutions using this technique right now. An example is a company called SAM.
Please share your thoughts and experiences in this exciting area in the comment section below. You will need to register. We appreciate your interest and expertise on EDN.
Steve Taranovich is a senior technical editor at EDN with 45 years of experience in the electronics industry.
- Archaeological Prospection of Northeastern Honduras with Airborne Mapping LIDAR, Juan Carlos Fernandez Diaz, William E. Carter, Ramesh L. Shrestha, Stephen J. Leisz, Christopher T. Fisher, Alicia M. González, Dan Thompson and Steve Elkins, IEEE 2014
- Digital Archaeology Field Recording in the 4th Dimension: ArchField C++, a 4D GIS for Digital Field Work, N.G. Smith, M. Howland, T.E. Levy, IEEE 2015
- “DEM generation from laser scanner data using adaptive TIN models,” International Archives of Photogrammetry and Remote Sensing, P. Axelsson, vol. XXXIII, pp. 85-92, 2000