At this time last year, I was buying brand new super-socks online for my long trip to Spain. I had a database full of hotel options, an international data plan on my phone, the perfect drop-down suitcase, and a new Canon T1i camera. Put simply, I had my sh-t together before United rolled it into a tight little ball and flung it at the fan. Given my travel history, one might think that while preparing for a second trip, I would try to simplify things… I didn’t.
Last time out, I took a ton of photos at architectural sites ranging from Barcelona to Portugal, and nearly everywhere in-between. Starting March 1st, I will be taking thousands of photos of just two sites: Calatrava la Nueva, and Montesa. These two sites were well covered in my first visit(s), but this next campaign is a totally different animal. This March, my wife Jocelyn and I are going to be using simple DSLRs to capture dense 3D point cloud data using a process known as photogrammetry. Rather than go into a lot of detail about the process here, I will just give you my favorite definition, and show you a few screenshots of photogrammetric point-clouds and models. Photogrammetry is best described as a reversal of the photographic process that captures the three-dimensional world in two dimensions.
The first image in this gallery is a Photogrammetry point cloud of Southwell Cathedral that was created by processing more than 1400 photos taken by my advisor, Lisa Reilly during two trips in 2007-2009. Each triangle outside the point cloud designates the location and focal length of one photo in the collection. The other images are described in more detail in the Photogrammetry section of this site.
By moving the camera around the site and shooting overlapping images, the newest photogrammetry software is capable of automatically matching common features between the photographs. Several stages later, the software triangulates the photos so that these features can be placed in 3D space as clusters of points. The process also corrects for perspective and barrel distortion in each of the individual photographs. This cluster of points is usually known as a “point cloud.” Until recently, large, expensive, and stationary laser scanners have been the most common tools for collecting this kind of data.
Laser scan of a nave bay at Southwell Cathedral. http://www.medievalarchitecture.org/
When the fall semester started, the only point clouds I had ever come into contact with were derived from laser scanning. I did some 3D modeling from laser scan data at IATH as part of the Southwell Minster project in 2009 that was later published in my advisor Lisa Reilly’s article “The Medieval Design Process at Southwell Minster” bound in Ashgate’s New Approaches to Medieval Architecture. When the fortress-monasteries of Montesa and Calatrava la Nueva emerged as my key case-studies I thought I would have to get my hands on a laser scanner in order to capture enough data from the ruined walls to build a faithful reconstruction of each complex. So when UVA’s Digital Media Lab purchased a new FARO Focus3D scanner, I was psyched. The downside was I realized this fall that there s a lot of red tape involved with borrowing a $30K instrument the size of a toaster and bringing it to Spain as a carry-on. In the end, the forms involved, and the lack of a secure place to store the scanner when I was not using it made me decide not to bring it afterall.
I wasn’t disheartened for long, because in photogrammetry, I found a cheaper, more mobile, less anxiety inducing data-capture option that would not only produce a 3D point cloud, but would yield a fully textured model at the end. After trying out a bunch of different software options, IATH purchased a copy of “PhotoModeler Scanner.” This software will be the primary package I will use for my photogrammetry work in Spain. I have also had a lot of success with Agisoft PhotoScan and the open-source VisualSFM. My introduction to the process came when Wayne Graham at the Scholars’ Lab showed me his first experiments with Autodesk’s 123D Catch. I would recommend this service to anyone who is just beginning to try photogrammetry, but once you get serious and you don’t want Autodesk to own your work, you should consider one of the previously mentioned packages like Agisoft PhotoScan or VisualSFM.
OK, so now I am a photogrammitist. All I need to do is shoot pictures with a consistent focal length, focus and exposure with proper overlap and the computer does the rest. What could be simpler!?
So why did I recently catch myself packing a new 6 ft. duffel bag with two painter’s poles, a remote LCD screen, a new tripod, a 7 ft. wide kite, 1000 ft. of string, five scattering laser-pointers and a bizarre looking metal contraption known as a “picavet?” The easy answer is that I can’t let something as lame as gravity keep me from getting the shots I need. Photogrammetry is more portable than laser scanning, so what am I doing taking pictures down here on the ground like a sucker!?
Go to www.flickr.com and search for “Pole Aerial Photography” and you will see some awesomely geeky hobbyists advise each other endlessly on how to get a camera on the end of a pole. The rigs they build are not very complex, but there is an obvious threshold when each one of them has to ask himself “Am I willing to convert my Honda into a 70 foot mast with wheels, or do I want to stay married?” In most cases they decide to stay with over the counter painter’s poles and some sort of digital or analog shutter release they can operate from the ground. While I never considered a mast, I logged enough hours staring at telescoping poles online that I did begin to wonder if “pole use” might send the wrong message on my CV.
Unfortunately for Jocelyn, Pole Aerial Photography (or PAP if you want to be short and gross about it) is just a gateway drug. If you want to meet a consistent group of basement-dwelling, newly divorced “tinkerers” just search “Kite Aerial Photography on Flikr. These “Kappers” start with a simple question just like the pole people: How do you attach a camera to a kite string so you can get higher altitude shots? If Kappers agree on anything, they all agree that a “Picavet” is the best answer to this question.
At its most basic level, this device has a cradle of strings that will adjust on the fly and level the camera that hangs beneath it. When in the hands of a Kapper, this thing gets pulleys to reduce friction, radio controlled servos that rotate and tilt the camera in any direction, video relays, and a flux capacitor. While I stopped at pulleys, I built the Picavet and cage from 30 or so pieces bought during a couple of epic “build it in the cart” trips to Lowes, and supplemented with some stuff from Amazon. This thing has sat on our coffee table or hung from a hook above the loveseat in our TV room for about a month. Jocelyn says I adjust or re-string the picavet so much I am making the poles jealous. She also says I spent a week somehow working the word “picavet” into most sentences.
Due to some Amazon shipping errors, I only received the kite last week, but our tests last weekend went very well. Thanks to Kelly Johnston and Chris Gist at the Scholars’ Lab for getting me up to speed on the kites. Kelly took a few pictures while we were out at Darden Towe Park that are worth seeing. The last image is the latest addition to the mix – a home-made kite reel I drilled out and built on Sunday.
The kite thing is not the end of my photogrammetry evangelism. Thanks to Louis Nelson’s Field Methods in Historic Preservation course, I am helping to teach digital capture methods along with Wayne Graham at the Scholar’s Lab, and Will Rourk at the Digital Media Lab. This is the kind of course that I would love to take if only I wasn’t teaching it. Louis Nelson is training the students on how to create traditional measured drawings, Wayne and I are showing them photogrammetry, and Will is demonstrating the FARO laser scanner. Louis found an excellent site 20 minutes outside of town called “Little Mountain” with 18th, 19th and 20th architectural features. I am heading up to the site on Thursday, so hopefully I will have some aerial shots for the next blog.
Jocelyn and I are leaving in just a couple of weeks, so there is a lot of preparation left and not a lot of time. Thanks to a check from Iberia Airlines, I have been able to replace a lot of the things that were in my lost suitcase from the first trip, but I am definitely not going to bring as many clothes on this trip. The really difficult part is going to be convincing Jocelyn to do the same. If there was any way that I could possibly make this trip without checking a bag, I would. Alas, the picavet is metal, there is no way in hell I am getting a 5ft telescoping pole into an overhead bin, and Jocelyn isn’t exactly the “backpacker” type. I’m not sure I have the cash, but I am considering some sort of tracking device for the long duffel bag that could be OK’d by the TSA. I will at least have a very durable tag on the bag this time. Wish me luck. I will need it.
On a personal note, I am excited to be able to show Jocelyn some of Spain, even if we are not necessarily going to the most interesting tourist spots in the country. I want to show Jocelyn Toledo on the way south to Calatrava, and I have never been to Valencia, so we should at least be able to see some cool stuff while the fortresses are closed. I have always thought of this as primarily a travel-blog, so there will be a few more posts in March outlining the adventures of Ed & Jocelyn in Spain. If you have been to the site before, you may notice that there have been some content changes also. All my digital work is getting collected into its own page, and I am trying to do a better job of wordpress-cleaning. I will be adding more examples of my work, including samples from my 6+ years at IATH, and whatever comes out of this last trip to Spain.
Finally, I was recently interviewed by the Chronicle for Higher Education for an article written by Stacey Patton titled “The Dissertation Can No Longer Be Defended.” The Chronicle’s version of what I do is outlined in a jump-page titled “Three Ph.D. Candidates Who Are Doing Digital Dissertations.”
If you are at an institution that can get behind the paywall, you can find the article at www.chronicle.com. If not, here are two links should work temporarily:
It was exciting to be part of it, but a little worrying as well. I try to do my best to make my work approachable in this blog, but I had difficulty finding the right level of terminology for the Chronicle’s intended audience. The lesson I learned is if I don’t have the best “low-tech” placeholder terms at the ready, only the bluntest things I say in the interview will be quoted. For instance, I am not surprised that my offhand comment about historians attempting to link medieval Spain with 9-11 became an even broader generalization about all of Christianity and Islam and led the profile. Still, the profile was fun, and I hope it leads people to this site where my obsessively controlled message can bore its way into more brains. Think about it. Don’t you have an odd itch to buy a kite all of a sudden?
<Rubs hands together and smiles.>