This paper is part of the 2017 3D Digital Documentation Summit.
Integrating 3D Digital Documentation Methods for Cultural Resource Identification and Documentation in a Heavily Forested Mountainous Landscape
Speaker 1: First I’d like to thank them for their presentation. We’ve done a number of — and Frank Lloyd Wright falling water and we can attest to Frank being challenged by the, to reinforce the [inaudible 00:00:18]. Because it’s a constant [inaudible 00:00:22]. Because there is a constant.
And also like their presentation, what we’re very interested in is the unseen, and trying to get at the less obvious, the less visible. Before I move on and know how fast to go, how many people here have used airborne lidar. Enough. Okay. How many people could care less about airborne lidar. Okay. Then I will spend a little time with it.
So the organization of my presentation is just the context issues with the methods being used, a couple of projects, examples, and future initiatives. So the context, then again I’ll just fly over that. A lot of what I’ve been seeing so far, I’ve been very excited about, but it seems to be from a kind of a heritage standing cultural resource perspective. Well I’m an archeologist, so mine is basically going to be focused on remains, ruins, traces, and absolutely nothing there that’s obvious. So with that as a perspective the focus is on landscape archeology. And something that we call landscape phenomenology, which is, the architects in the room would know that, and it’s basically trying to understand how people perceive and interact with their environments.
So where we’re working, we’re working in southern West Virginia in the mountain top mining region. So depending on what you’ve read or haven’t read about West Virginia, the anthropocene, which is basically just an explanation of the man made geologic construct we’re in right now, is alive and well in West Virginia. And basically it’s alive and well in Mt Tucker Lousiville mining.
So we’re working in the new river gorge, which is a unit of the national parks service, and a couple of other areas. So with an explanation there, the new river was some of the earliest coal mining in North America, and it is currently an area that is managed by the national park service. Primarily as a natural resource and cultural heritage park.
Then we’re also working in the Appalachian geo park which is a recent initiative which is basically right now the UNESCO world heritage group has no geo parks in the united states. There are like, two in Canada, and that’s it. And there are actually two efforts underway, one in Michigan and one in the Appalachians. So what is a geo park. It’s basically where the geologic geography has driven the culture that’s occurred there. Whether it’s back to native Americans or moving forward. And then the third area, which was down south, was Gary Hollow, which is an area that we’ve studied a lot in McDowell county west Virginia and it’s an area that’s been judged to be appropriate for world heritage status, because again, for over 50 years it was the largest coal mine in the world.
So the region. The region is kind of very rugged mountainous terrain, very heavily vegetated. And that has implications, because things left in the landscape will vanish within 20 to 25 years, and then there’s just basically hardly any trace. Very rocky, very poor soils. Doing any subsurface work is never gonna happen, and traditional field work methods have proven to be very difficult. And again it’s just because of how things vanish relatively rapidly.
So the extensive multi-dimensional effort that we’ve embarked on, and again we were doing all that early pre-technology field work. We were out there with transits and levels and tape measures and things like that , and it was just a nightmare. So the data collection, it’s been a fairly significant effort, so we’ve collected airborne lidar over the entire area, which was basically right around 6 million acres. We have higher density of multi-temporal lidar, over about a half a million acres, color for color infrared imagery over the entire area. Terrestrial lidar, and we’re doing a lot now with drone based photogrammetry.
So we have a number of technical issues we were actually interested in and one of them had to do with the research interest of mine and that is using anomaly detection methods to find things in the landscape, that the rules of the landscape tell they ought to not be there. Because more often than not in the Appalachians, that has some sort of man-based influence in them.
So what we’re doing is we’re integrating these technologies with other data, and what we’re integrating are basically the beginning point; terrestrial and airborne lidar. And so what we can see is an area in which we’ve done the terrestrial laser scanning and the much larger landscape of where we have the airborne lidar. And again this gets into the kind of artifacts in the landscape that I’m interested in are things like those running remains and traces rather than heritage buildings and things like that. So there are the trees and what everything looks like now.
And this is kind of the mapping we did the first time through, it was incomplete. It was not terribly precise. Nor horribly accurate. So airborne lidar is the base for data collection, and what we can see here.. and this also is kind of an add-on to my interests, are if the post modern mining had not occurred in that region, this would have been one of the most historically significant small towns in the Appalachians. But it was completely obliterated. So one area that we’re actually working with UNESCO on is the impact of mountain top removal mining in the Appalachians. Should that be a world heritage consideration, just given it’s significance. So again, that’s just the kind of landscape in which we’re operating in, so you can where the landscape as a backdrop to all of this pretty significant.
So terrestrial lidar in the integration of these things, so what we can see is a small mountain community and we can see where it fits into the larger landscape. So the issues with airborne lidar are canopy is an issue, ten shots per meter squared. And with terrestrial we hear about hundreds of thousands of points per meter squared. With airborne it’s much less than that. What’s important there is we work our way back and we begin with the density that we want to hit the ground, and then we do kind of an analysis back and get kind of mid-ground and then our overall, how much lidar data we want to collect in order to get that configuration on the ground. And what’s important is lidar processing is really an art, it’s not a science. And what that has allowed us to do was we were able to bypass consultants and contractors and engineering firms because we have our own airborne lidar unit.
So we were able to gather raw data, which most organizations are not able to do, and why I said anomaly detection hunting for kind of human anomalies in the landscape, is proven to be really difficult, because anomalies are actually an artifact of airborne lidar data collection. Whether it’s due to climate or rain or just he laser itself, that data that’s collected you’re collecting a lot of data that you actually don’t want and it’s actually confounding the issue.
So what’s important there is, again if a lot of us if we’ve looked at airborne lidar data, and if you’ve generated your hill shade in art map for example, and you’ve defaulted to the defaulted art map, don’t do that because that’s a distant weighted, use Creegan for that because you’re going to accentuate the anomalies rather than hide the anomalies like the defaulted art maps do. So some lessons we’ve learned, automating anomaly detection has proven to be pretty impossible. I’d call it difficult. Lidar data may be processed to be a peer consistent by consultants and in fact it’s not. So basically what you’re doing is masking the quality of the data is what most engineering firms actually do. [inaudible 00:09:54]
And it was just things like remains of rocks, you can’t tell the difference automated, so. So we also some other devices. We have a drone that we’ve built from this morning you saw the gigapan, and we have this wonderful thing over on the far right and if you were looking at that and saying what should that thing be called, well it’s actually called a wonder pole. And what we do is we put cameras up on that, we’ve had a laser up on that, we’ve had a hyper spectral remote sensor up on that. And it just works like a, it’s awesome, it’s only about 20 to 25 pounds and we can lug it everywhere. So it actually is maybe the 8th wonder of the laser scanning, and collection for archeology.
This was important, if we’re interested in this meshing of airborne terrestrial data, how do we actually do that, and what are the kind of results that we get. Well we collect the airborne with what’s called RPK GPS, which is basically real time schematic, so it’s all over the place. We don’t use that for the terrestrial, we use what’s called static occupancy, where we set ground control points, and just almost make ourselves at home on those points, for often half an hour per point.
Then what we do is we register these two things, and then the example over here on the right was every time we’d do a project where we’re going to mesh these two we will take some swaths and we compare them and we mash them up. So in this example it’s a one meter wide swath. The airborne data we can generally get within 6 cm of any ground control. Anywhere in the Appalachians. Which for our work is really pretty good because we’re doing, as the surveyors tell us all the time at the end of the day, it’s just archeology. But they were able to get our terrestrial data within plus or minus 2 cm of that airborne data, which we feel for the kind of work we do is great. And then I also like that flag because it’s the place where we did this was over in the Montgomery county Maryland, and that in fact, this little structure here is in fact uncle tom’s cabin. Say what? Well it actually is. That’s what’s the inspiration for Harriet Beecher Stowe’s novel was that little element right there.
I can get to that exact location, it’s actually privately owned.
So we collect imagery every time that we fly lidar and we’re collecting more and more and retrieve with the drone. I’ll go over that. Some of what we’ve found for being very useful is not just the extended arc map hill shape, but to generate multiple hill shades, run principal components on that, and then you’d get all of those hill shades in one hill shape. So tiny little things basically pop from the lidar data at that point. So whenever we’re out there, even if it’s not a part of the mission, we’ve scanned things. That’s an early 19th century corn house. That’s a mid 19th century farm stead. So what we also collect, is we collect the all points and the bare earth. And that’s because we’re interested in what it looks like and what we can discern on each.
For example, in the all points that’s the coal loading building, in the bare earth that’s the location of the coal loading building. There’s a conveyor going up, and early 20th century conveyor over on the bare earth is the location of where that conveyor was excavated for. And these are just some of those objects that we just looked at on the lidar with some photos and then a laser scan of its walls. The mine building which the parks service reconstructed unfortunately, and then the laser scans of that park service building.
I’m a strong believer of just leaving things in the ground and we don’t have to Walt Disney everything.
Our work in the Appalachians is farmsteads, the rural industrial landscape. Appalachian mining and mountain communities and adaptive reuse in recreation and opportunities.
So an example again, so this is our terrestrial laser day of getting the four or five farm buildings. There’s that data overlayed on our airborne data, and as we zoom in, we can see within about a 2cm agreement between the airborne and the terrestrial there’s that as we move in with the airborne data in the background. And that’s some of our final models. And obvious from the green spheres, we do use targets. And I just don’t think we’re smart enough yet to use them. And then we’re also integrating kind of interior and exterior, where we’re looking at the basement of one of those buildings.
So drones. We are now, because our airborne laser was about nine years old, and it’s an Optech 3100, and Optech wanted this last year $140,000 in annual maintenance, we decided to put that rascal to pasture. So we’re now flying drones. And what we’re interested in is combining airborne lidar and the structure for motion, or the uncontrolled photogrammetry and structure for motion with our terrestrial lidar.
So this is a digital surface model from a drone, and what we were interested in in this area here, because this is an area that had been very heavily vandalized because of native american remains were discovered there about 50 years ago. And there is the elevation model where we can actually see the results of a lot of the digging on that vandalism which was not obvious.
So we’ve gotten pretty good at these structure for motion things. This is Harper’s Ferry West Virginia, and although it’s not in that study era, this was the very first time we did it. I was able to generate that model, never having looked at the manual , having not idea what I was doing at all. But it’s really not bad. We were able to get in there and measure 6″ steps and things like that, so. The structure for motion holds a lot of promise. Someone in the morning group said, how does the precision line up with laser scanning. Laser scanning is far and away far superior in terms of both accuracy and precision.
So then what we were trying to do, we’ve encountered some problems rectifying our structure from motion kind of natural force with the data from the airborne lidar, and we’re not sure why, other than you can see when we go out to the corner of an area here we’re beginning to get some drift. And that drift turned out to be about half a meter or so, so we are still working at trying to develop this, and I would suspect that it was a matter of the number of ground control points.
So what we’re currently doing is we’re supporting heritage documentation, and Gary Hollow again is this area that’s basically potential UNESCO world heritage. So what we’ve actually been trying to do, this is just a small part of the time lapse that we’ve built up for this area, so we have an area that’s what these towns looked like back in the day. So there’s what kind of culture looks like in the all points. So what we were actually interested in was measuring the heights and the configuration of the roof lines of the buildings, trying to do a categorization of these early 20th century mining structures. And here’s an example of that, back at the same site.
So what do we actually do with that, we’ve found that that was one of the greatest concentrations of turn of the 20th century mining community houses still in existence anywhere in the united states.
So then we said well we’ll probably want to go out and terrestrial laser that, so that’s what we did, and that’s what this community looks like.
We’re also trying to identify locations of potential remains of mining among residential areas, and that’s because part of what we’re doing is trying to document the impact of mountain top removal or mining since the 1970’s on the loss of historic cultural resources that ought to be important, but are lost in the Appalachians.
So that’s our town of Gary study area. That’s a 1909 US GIS map. So I condensed this way down.
That’s an earlier DEM.
There’s an area where we know back in 1909 there were three coal mines. Then we built up a time lapse of that.
And again, three other coal mines.
All of those would have been potentially significant cultural resources if they were still in existence.
And this is my poster child for lost resource. And that’s just kind of what that landscape looks like.
So we’re also documenting things that we just find and it just worked out amazingly that that scan of that coke oven and that coke oven are actually the same coke ovens. Now I can say that was a planned field investigation now. We had the numbers on the cook ovens tied to the historic ovens, and there was a plaque right under there with the number of that cook oven. So that was just fortuitous.
And because of a personal interest of mine in landscape disturbance, I’ve an ongoing study in the Appalachians of iron furnaces. And so there’s an example. There’s kind of the rough pre-scan of the iron furnace. And we’re able to measure volumes and voids and everything like that and there’s what the data looks like.
There we are inside the furnace. And that’s the detail that we tried to develop. An aerial view and there’s that same iron furnace in the airborne data. And then we’re also doing some things with the historic bridges and tunnels. I like that graphic because basically it’s a historic railroad tunnel going right through the mountain from the airborne data.
So what were’ doing with that data on the tunnels and the railroad bridges, well right now in the Appalachians there’s a movement underway for rail trails, and if you have a rail trial in the Appalachians, per linear mile there are going to be an amazing number of tunnels and bridges. So what are we doing? We’re working on refined documentation of features and sights that have been previously visited, often by us, and found out with the current technologies that we have how bad our early field work actually was. We’re working, progressing on Gary, identifying the status of the historic mine features and in Gary, we’re also historic neighborhoods with significant collections of historic structures and we’re prospecting for new features.
So that was it in a nutshell.
This paper will discuss a number of different technologies that have proven applicable to cultural resource prospecting and documentation – including airborne Lidar, terrestrial Lidar, structure from motion photography, and drone based remote sensing. More importantly, the paper will discuss our research in integrating these technologies to provide fused data multi-scale documentation for features in a particularly difficult landscape for such documentation – heavy deciduous forests, steep topography, and a history of what has often proven to be poor or incomplete documentation. For example, the Natural Resource Analysis Lab at WVU using its ALTM 3100 acquired over five mission acres of dense airborne Lidar data for what were the most historic mined areas in West Virginia. Where appropriate, we have integrated this data with terrestrial Lidar to capture specific detailed data for features of special historic interest within the broader data from the aerial sensor. Using static occupancy GPS for both the airborne and terrestrial data, we have been able to link these two different data sets often single centimeter precision or better. This has also proven to be applicable to Lidar/drone imagery; drone imagery / terrestrial Lidar; and structure from motion point cloud data / terrestrial point cloud data.
This work has focused in and in the locale of the New River Gorge National River (USDI-NPS) and Gary Hollow, which is located in McDowell County WV, which for over fifty years was the largest underground coal mine in the world. The presentation will also describe work completed along the Ohio River and areas in Northeastern West Virginia in the high mountains of the state. Mining towns and features, historic highways, frontier settlements and work on Civil War battlefields will be included. The work is ongoing under the umbrella of a project initiative titled “Appalachian Heritage – Remains, Ruins, Traces and Structures”. The work is also focused on providing linkages between the public history of sites, locales, and regions which are often relatively well documented in terms of oral and written histories, with past and current site conditions with specific landscapes and remains which are often very poorly documented. The work is also now supporting early documentation efforts for a proposed Appalachian Geopark. Geoparks are a program of the United Nations UNESCO program with Geoparks integrating environmental importance and uniqueness with cultural and historic resources. This would be the first Geopark in the United States while there are over a hundred Geoparks throughout the rest of the world.
Charles Yuill is Associate Professor in Landscape Architecture at West Virginia University where his research of over twenty five years has focused on cultural resource documentation and management, industrial archaeology, and 3d documentation of environmental conditions. The opportunity to work with early terrestrial Lidar data (using one of the first Cyrax scanners) of medieval features in Scotland allowed him to realize the opportunities of such documentation- with the work now focusing heavily on industrial heritage in the Appalachians.
Peter Butler’s work at WVU focuses on heritage and cultural resource management, as well as community design outreach and engagement. He has worked in small Appalachians communities focusing on the roles that heritage tourism can provide in the recovery of these small rural communities.