Keystone Aerial Surveys, Inc. recently purchased an UltraCam Falcon, the “prime” model with the larger across track footprint of 17,310 pixels, with a 100mm focal length lens. [See press release.] This is the third such system in 12 months for Keystone who now owns a total of eight UltraCam systems that also includes three UltraCam Eagles and two UltraCamX systems, making the Philadelphia based company one of the largest UltraCam operators worldwide.
I recently had an opportunity to sit down with John Schmitt and David Day, President and Vice President (respectively) of Keystone, to talk to them about their operations and success that has taken them from a once entirely film-based operation to a leader in digital data acquisition in the U.S. Below is the transcript from that interview.
– Jerry Skaw, Microsoft UltraCam Team
Thanks for taking time to talk. For those who might not be familiar with Keystone, can you provide some background information?
SCHMITT: Keystone was started in 1963 by Joe Mullen and we just passed our 5th decade in business. The company started with five people. Shortly afterward Gill Mallinckrodt joined and became president and owner. Gil was followed by Ken and Mary Potter. When I started in the early days of Keystone, back in 1987, we had roughly 20-25 people and six aircraft that were mostly Cessna 320s, the Cessna Skynight. At that point in time we were just doing film acquisition, eventually flying as many as nine film cameras. Most of our work was generally from West Virginia through the northeast [USA] and was very seasonal: spring and fall, with spring being the primary season and fall the lighter season. It was primarily for small, photogrammetric engineering size projects. At that time there weren’t really a lot of very large blocks. Some of the projects we would fly in the summer were projects similar to the National Aerial Photography Program (NAPP) for the USGS. We still utilize the nine film sensors but our goal is to replace them with the equivalent number, if not more, in digital systems as we move forward into the future.
What year did KAS purchase its first digital sensor?
DAY: We bought the UltraCamD in November of 2005 and started using it in 2006.
Tell me about that decision. I talk to film flyers still today who struggle to make the leap. Either they are unsure where the work will come from to keep the camera busy or they have trouble getting past differences in technology.
SCHMITT: We knew the age of digital was coming, that we would have to go digital eventually and didn’t want to be one of the last few companies to get into digital. So we thought we would be an early adopter.
DAY: We could see it coming. I liken it to vinyl versus CD. Sure, some people still say vinyl is better but CD took over. Obviously new music is all digital, not even CD based, but we knew the industry was going in that direction and we didn’t want to be left holding a bunch of vinyl records.
SCHMITT: We actually had to do a lot of education on our part with the client, flying projects for free, so that they could compare their film projects with all their established ground control and set up. We’d fly the equivalent with digital so they could compare apples to apples and they could see what they were getting not only with the clarity of the images but more importantly the accuracies.
DAY: Once you go digital, you don’t really go back. We’ve seen a lot of people—sometimes new employees—come in with the attitude that film is much better and when we actually put them in front of digital and have them use it. They realize they prefer digital in the end and never want to go back. There’s a lot of preconceptions and misunderstandings of what to do with digital imagery. But once you can get it into people’s hands and get them started really using it and understanding how to use it in a different way, taking advantage of the computing technology, then they realize the advantages.
Dave, I want to ask you to expound on that a little because it touches on a topic that I have discussed with companies who still have a traditional back office workflow and who are concerned about vertical accuracies because they say they don’t see the vertical exaggeration with digital imagery as they do with film because of base-to-height issues.
DAY: Yes, we’ve had clients who have heard that film inventories are becoming limited and have contacted us with worries about base-to-height ratio and other differences. I explain to them that it is true that if you are doing purely visual compilation and you need to rely on your eye, you may need to use different methods. The issue is in those cases is that film may be slightly more accurate but if you slightly over-engineer the project—meaning you fly at a slightly lower elevation to get a better GSD [ground sample distance] and you take full advantage of the digital process of auto correlation—the base-to-height ratio is not nearly as important. You are using the redundancy of the computer-aided measurements to overcome and supersede the manual method. Then you compare the time it takes, if the digital process requires a quarter of the time needed to complete the job manually, and its 98% as accurate, isn’t that a viable trade-off?
Many of the current film superiority ideas were developed when digital cameras and software packages where in their infancy. With the maturity of the technology have come advancements that render mute past arguments.
SCHMITT: The base-to-height ratio is essentially in the eye of the beholder. Since we’re an acquisition company and we provide flights for so many customers, we get to see a lot of the flight plans come over. Sometimes they’re the same flight plan, the same area, but people plan it differently. They’re idea of what is an accurate specification and resolution changes depending on the company. So we might get the same project in from three different companies and there might be three different GSDs. And it’s all dependent on what they feel comfortable with. We also have customers we have flown certain projects for and who are trying to achieve 1 or 2 foot contours and over time, as they have been using digital and becoming comfortable with it, they’re actually going to a coarser GSD as the technologies, cameras and software advance.
Keystone started with film, made the decision to go digital, and has since consistently increased your digital capacity. You’re doing more and more digital flying—presumably now flying more with digital than film. What is the current state of film in your opinion and what do you think the future holds for it?
DAY: Actually this past year for Keystone was the first year the pure flight hours for digital projects were greater than for film projects. I think we were at 60/40 last year. So it has flipped and the state of film right now is very much in question. We’re hearing from some of the manufacturers that availability is getting difficult to assure, especially for the color film.
SCHMITT: All the film from the manufacturers is going away slowly but surely. The folks in the US are the main purchasers of the film. In Europe and other continents, companies just decided “we’re going to go digital” and they just flipped the switch and the next day they were all digital. For us, with the demand from our customers for film and with so many planes and cameras at the time, it would have been a foolish decision to abandon film and our customer base. But the manufacturing of film is going away. I would think maybe 2-3 years, if the manufacturers don’t decide to reinvest in making film again, it will be the end of an era. How many film projects are there left? What we’ve noticed is that some of the film-based projects that we’re flying for government purposes are slowly starting to transition to digital by saying “let’s try 10% of the project in digital.” Then the next year “How about 20%?” So we see the transitioning. If you look at the USDA—all of the national forests projects—they have all switched from film to digital. There really aren’t that many programs that are looking for large-scale film.
Let’s talk about the PASCO acquisition in 2011. What has that done for the business?
SCHMITT: The financial backing has certainly enabled us to plan a more aggressive transition to digital that we could not have done as Keystone alone.
DAY: It is nice to have an investor and now an owner who is in the geospatial industry. When we make requests for assets and to expand our capabilities, they understand why. They understand the marketplace and the geospatial world. They understand the capital expenditures that are necessary to maintain a state-of the-art company. On a day-to-day basis, it’s also very nice that—whether it is culturally or just the way PASCO operates—they trust us to make the right decisions.
Has the way you conduct business changed?
DAY: It’s been pretty seamless. The average employee here has not seen any real change in his or her day-to-day life because of the PASCO purchase. Keystone is run basically the way it was run before the purchase. That’s a testament to how well it was run previously but also the trust factor of the ownership in the current Keystone management team that we can continue to run things on our own and make good decisions. What’s also been beneficial in having a parent company in the industry and a worldwide group of PASCO companies, is the sharing of information, of ideas, and sometimes even resources.
The business model has changed a little but only because we are focused on providing more growth. In the past we may have only been able to provide a certain level of support to our clients. Now, with a larger company in the background, we’re able to meet any demand. Our customers can go after projects that they didn’t think possible before because they know they have Keystone and PASCO behind them. Whatever you ask of us—if it’s a big enough project for a long enough time—we’ll be able to put together the resources to do the job. Now virtually, nothing is off the table whereas before we may have been limited in our capabilities.
SCHMITT: Added to that, when we were flying film we were limited on the products we could supply. We have now been able to branch off and provide more products and services through digital acquisition that enhance our deliverables. This, in turn, is a benefit not only for the larger companies but is a major benefit for some of the smaller firms who want to be able to take on the work but who just can’t because they have limited resources and now know they can team with us on these projects and be very successful.
By increasing digital cameras, it has obviously increased our opportunities. Not just with our customers but also with our competitors. We can now actually work much more closely with competitors on projects of any size. Sometimes they win the projects and we actually acquire the data along with them.
The number of cameras has certainly increased. While many flyers have held off making investments in additional acquisition tools in just recent years, citing the post-recession economy, Keystone has purchased 3 UltraCam Eagles and 3 UltraCam Falcon Prime systems. What is driving your success?
SCHMITT: We’re staying true to our goal and that’s to be an all-digital shop in the future. When certain projects come out that are fairly large, the end client has an opportunity to either pull a consortium together of all different sensors from all different providers, or they have an opportunity to work with somebody with multiple cameras. So I think we’ve been able to find success by working with larger companies and being their sole provider of imagery rather than having a partnership and Keystone just being a single part of the consortium. And I think Keystone’s reputation of being able to get the job done in an efficient manner has added to our success.
Does operating the UltraCam provide you with opportunities, directly or through partnerships, that you might not otherwise have if you were operating a competitive camera?
SCHMITT: I know some companies don’t want to marry sensors together, even though they can. We’ve seen projects where we knew it was possible to use other sensors but they didn’t want to mix cameras and they said “if you can commit multiple UltraCams to this project, we’ll use Keystone exclusively.”
After purchasing three UltraCam Eagles, in the last year Keystone has added the UltraCam Falcon Prime to its arsenal. Tell me about that decision. Is that because you are seeing a different kind of work that the UCFp is better suited for? What’s going on there?
SCHMITT: I think each project has its own special needs. For the larger, big blocks (counties, state-wide work), the UltraCam Eagles are great. When we fly low scale collections, the UltraCam Falcon provides a faster cycle time and a nice sharp image.
DAY: I agree. Because of the cycle rate and that we opted for the 100mm lens, the Falcon gives us more flexibility for the smaller jobs. But then because we specifically bought the Falcon Prime [the larger footprint version of the Falcon with 17,310 pixels across the flight path], the systems are also flexible enough for the county work. So the Falcon is getting close to that 1-to-1 replacement for a film 9” x 9”.
We [UltraCam team] boast about efficiencies because of the footprint of the cameras reducing flight lines and mission time but there are also production efficiencies through the UltraMap software. Dave care to comment about that?
DAY: The enhancements in the post processing software are great for a company like ours where we have a lot of different clients, all asking for their job first. The tools are helping us balance that now. And it is high-level distributed processing that you can manage from anywhere in the world and get constant updates, and target the machines—and even clusters of machine—to take best advantage of the hardware you have. We’re happy with the updates we’re getting from Microsoft. It’s been a great help. I’ve heard through the grapevine that other manufacturers don’t spend as much time updating their software as Microsoft/Vexcel does, so it is just another selling point that keeps us coming back.
What do you think of the new subscription model?
DAY: I think it’s a good idea. It will help us spool up for busy seasons or large projects. And it allows us to consider going after further post-processing products like digital elevation models and digital surface models without having to make that initial large investment. We can find a project that will pay for the subscription period and determine if there will be more work to make the full investment. I’m a big fan of software subscriptions, buying what you need and not buying something permanent that you may not need. Software as a service and on demand with faster internet and globalization has made this the model of choice for consumers.
Keystone also operates LiDAR devices. What do you think of the concept of point clouds from photogrammetry?
SCHMITT: They each have their own individual positives and negatives. With LIDAR you can derive elevations and see under trees. With photography, it is a bit of a hard to measure near vegetation but in open flat lands and roads, the point clouds are tremendous. I can see software in the future improving so that they will be able to model the trees a little bit better. One thing we do notice when we incorporate the imagery-derived point clouds with ground control on hard surfaces we can get better accuracies than can probably be achieved with lidar – especially in horizontal.
DAY: Lidar’s advantages are that you can collect more frequently because you can fly at night, under clouds, and measurement below the tree canopy, and the vertical accuracy is high. Now, everyone knows that with lidar’s horizontal accuracy you’re only as good as your GPS because horizontal control for lidar is impractical. It’s also very hard to match the density of imagery-derived point clouds. The best lidar, maybe, is 50 points per square meter—the latest models might be able to do that—whereas you can easily get 300 points per square meter from an image. But just based on the mathematics of sight lines and rays, there are more artifacts than you would get from just getting laser returns. The advantages of the imagery derived point cloud is that you can cover a really large area because you can use an UltraCam Eagle for the collection so you have less flight lines and a point cloud is a derivative product from something you may be doing already. You may want the imagery to produce orthos, engineering work, or any kind of planimetrics. Simply adding a little tighter spacing to the flight lines, you now have another product that you can maybe resell to someone or plug into your GIS (that are all going 3D) and get more spatial information than you ever could before. So as John said, each has its own benefits and depending on your need one will be better than the other.
SCHMITT: And most of the time people marry a camera system to a lidar. You need to be able to see a picture to pick out certain things with the lidar. So most of them are usually paired with a small to medium format camera. Here with the DSMs derived from imagery, you have a high-resolution image and a high-density imagery derived point cloud.
DAY: A point cloud that is perfectly registered to the imagery.
Keystone has been at this a long time and you both have been there for a good part of its history. Does the market/industry/profession look much different today?
SCHMITT: I would just say that the digital cameras have opened the door to fly more projects around the US. Back in the early days we operated in a specific geographic area of the northeast US. With our different offices, and the services we can now offer all through the US, we see that opening the doors to teaming arrangements and new clients, and spinoffs from companies that fail into new companies and it’s increasing our work line.
DAY: Satellite and internet imagery providers have changed the dynamic. We’re offering higher resolution, higher accuracy products than can be purchased from satellite or provided for free. We’re flying a lower and performing different types of collections than when I first started, certainly.
SCHMITT: It seems nowadays the projects are getting larger in size. Instead of small engineering projects, we’re flying big blocks, multiple counties if not states.
What are your thoughts about where the market is heading. Do you see changes coming?
DAY: We’re going to continue going lower and lower. We used to fly at 40K-50K feet with a Lear jet. That was gone as soon as satellites came along. We used to fly at 20-30,000 feet for 1 meter GSD imagery. Obviously that’s gone. Now, satellites can provide 30cm data so a bit of our 30cm business will go away. So we’re going to be flying more at 15-20cm. And more at 2cm if the UAV market increases and more people get their hands on imagery and realize they can get it from airplanes with the Falcon.
What kind of advice would you give other flyers, in particular those who have not yet moved into digital acquisitions?
DAY: Going back to the markets, there are small companies (1-5 employees) all over the country who are still stuck in film and haven’t made the investments in their back room or in their cameras – depending on their business. As film truly goes away—I think we are in the very tail end, the last 1 or 2 years—they are going to be where John said. They are going to be forced to rethink their business. Do they want to be in this business? Where will they come up with the capital? That is going to lead to mergers and some people not surviving and some companies joining together with existing companies or forming new companies in order to build up that capital/loans in order to invest. I think there will be a large consolidation. Those micro companies will go away and there will be just a few remaining companies, similar to hardware stores, banking, etc. For people who haven’t gone digital, flying-wise, they need to start planning that investment or start planning who they are going to team with and eventually merge with when film does not become available anymore. Because we are in that tail end of the bell curve and it’s time to get in or find somebody who can get help you get in.
SCHMITT: The end of an era.
Speaking of an end of an era: happily for them but sadly for the rest of us, Ken and Mary retired. Which means new positions for you two. How has life changed for the two of you?
SCHMITT: We’re adapting to the changes and new positions. PASCO has a great deal of trust in us. Keystone is poised for one of the busiest years we’ve ever had. We look forward to taking on all the challenges that are ahead and leading Keystone into its next era.
DAY: My answer that I tell everyone when they ask this question is that Ken and Mary did a pretty good job of handing off their duties all along so for John and [me] not much has really changed. Maybe 15% of our jobs have changed in the last 6 months. But leading up to that we were doing a lot of the day-to-day and planning for Keystone, so they did a good job of preparing us.