This month, Waypoint chatted with former geologist and commercial-pilot-turned-educator Scott Winslow. A remote sensing and GIS mapping specialist, Scott now uses his aviation background to conduct research and educate future GIS specialists in low-altitude remote sensing.
Please tell us a little bit about yourself.
I worked for 31 years in the private sector as an engineering geologist. A career change in 2014 led me here to California State University in Long Beach, California, where I currently serve as a lecturer and GIS lab manager for our university’s geography department.
Could you give us brief overview of your department?
Our geography department basically has two specialties, namely physical geography and human geography. I’m on the physical geography side, which encompasses remote sensing and GIS. We do things like mapping and cartography. We’re the part of the geography department that is actively using senseFly equipment and rotorcraft sUAS.
Which drones do you fly?
Right now we have three senseFly drones, the first of which was the eBee SQ we purchased about five years ago. That one is on a project site we’re working on right now in Mali, West Africa, where our faculty and graduate students trained local villagers and researchers how to fly [the drone] and collect vegetation data.
Our local colleagues in Mali are collecting multispectral imagery for us on a regular basis and sending the data back for digital processing. I am happy to say that, after a mishap last July disabled the drone, the eBee SQ is back up and flying over the West African savanna thanks to the quick response of your repair team here in the US.
Any other senseFly drones?
We have the eBee Plus, which has the original senseFly S.O.D.A. camera in it as well as the MicaSense RedEdge camera. After gaining some experience with the SQ, we decided to purchase the eBee Plus in August 2018. This drone has been our workhouse for the past year. I trust that drone implicitly. We had some little bumps in the beginning, some glitches and so forth. However, these kinds of field issues are now few and far between. These days, when we arrive at a study site, we can typically have the eBee in the air within 15 minutes.
The support I’ve gotten from senseFly… is unlike any of the other equipment suppliers we use.
This past March we took delivery of a new eBee X. The purchase of this drone was driven mostly by the fact that we needed the RTK capability for the large study areas we are working in. One such site is an 800-acre ecological reserve located in the western foothills of the Sierra Nevada mountains of California. This site is characterized by rugged topography, hot summertime temperatures and a lot of well-fed rattle snakes. Not exactly the best environment to send students out to set ground control points.
With the RTK-enabled eBee X and the S.O.D.A. 3D camera payload, we’ve generated a fantastic DEM of our project site that rivals any published LiDAR dataset for this area in terms of accuracy. We’ve also flown it a number of times with our Sequoia camera and generated some decent multispectral mosaics. We’re in the process of getting a few bugs worked out with the help of Daniel, one of your customer service reps, and I think we’ll be back up in the air with the Sequoia within the next month.
Drones are amazing technology, but sometimes things can and do go wrong. Obviously, it’s frustrating when that happens. What has your experience with senseFly been?
I must tell you, the support I’ve gotten from senseFly, particularly Daniel, is unlike any of the other equipment suppliers we use. We use a lot of Trimble equipment; we use software from several different companies, and nobody is as responsive as you guys are. The unique thing about Daniel is that, although his main goal is to get our problems solved as quickly as possible, he’s also truly curious about why these things are going wrong and seems fascinated with the whole process of getting these issues resolved.
Switching gears, how did you accomplish your work prior to using drones?
I started working in this department as a grad student back in 2009. That was when we first started using commercial-grade, fixed-wing UAVs for aerial imaging. I think that we were kind of on the cutting edge as far as the universities go. Before that we were strictly using satellite imagery and aerial photography like most people. And you know, satellite imagery is not cheap, especially when you task a satellite to capture custom imagery of a particular study area.
Back in 2008 and 2009, other faculty from our geography and anthropology departments had a project on Easter Island that required high-resolution aerial imagery. For this project, we purchased a Trimble UX5 and became one of the very first users of this early fixed-wing commercial UAV. Although much more cumbersome to operate than a modern autonomous UAV like the eBee, this drone with its Sony true-color and infrared camera payloads served us well and allowed our faculty and students to image a good portion of the island.
When we arrive at a study site, we can typically have the eBee in the air within 15 minutes.
That was the beginning of our interest in using UAVs to collect custom aerial imagery. We subsequently invested in some small multicopters. I did my master’s thesis using one with just a point-and-shoot Olympus camera that I taped to the bottom of the thing. This quadcopter had a flight time of about four minutes, so there was a lot of battery changing. But we were able to image a one-square-mile site in 39 flights, and the image mosaic came out surprisingly well considering how crude the rig was.
What made you start exploring alternative technologies, such as fixed-wing drones?
A number of years after purchasing the UX5, we bought our first senseFly product, the eBee SQ. We used it a little bit locally, then sent it to Africa. We then bought some other multicopters and were flying a RedEdge on that and it was pretty good, but it got a little complicated mixing and matching equipment.
Why was that?
For one thing, the integration of the camera to the UAV platform for those models was not nearly as deep as senseFly’s. With senseFly, I get spoiled because everything is all in one package—you don’t have to have three apps to juggle at the same time.
So, after that we started training students how to use all this technology, and we started getting more local project sites that were well-suited for this type of imaging. We acquired additional software that allows us to generate very dense point clouds and perform automated land cover classification. Taking all of this know-how to the field, we’ve recently embarked on a project in Central California. It’s shaping up to be a long-term study of changes in vegetation cover over time.
What would you say is the biggest benefit you get from using drone technology?
The number one thing is the cost effectiveness. Yes, it’s a considerable investment to buy a piece of equipment like [a drone], but if you need to task a commercial satellite to get a custom image of a project site, it’s going to be $3,000 just to get that one image. And if you want to get multiple images during the year, well, you’ve just bought yourself an eBee.
And let’s not forget the image quality. When you’re talking about off-the-shelf satellite imagery, we take what we can get. Are there going to be clouds in the scene? Well, I don’t know, that’s kind of a roll of the dice. Is it going to be hazy the day when the satellite collects the image? Is the sun going to be at the right angle? With the eBee, we can control all of that. We can go out and we can fly it when the conditions are just right – whatever month, day and hour we choose to fly it.
Why did you choose fixed-wing drones as your primary drone technology?
The thing is, there are specific applications that require the use of a fixed-wing platform, and other applications that require other drones. You really need to have both. When you want to cover large areas, though, you can’t do that with the rotorcraft because the flight endurance just isn’t there. For example, the site that we’re working on actively right now is 800 acres, so to cover that with a quadcopter would take a number of days.
The other thing is the topography; most of our study sites have very rugged topography. With our senseFly drones, the images we capture are going to have consistent spatial resolution because the drone is going to maintain flying height above the terrain. The rotorcraft we currently own don’t have this capability. You have to constantly adjust the flying height in order to maintain clearance with the ground not only to keep from crashing, but also to keep the spatial resolution of the imagery consistent.
With senseFly, I get spoiled because everything is all in one package.
What does a typical workflow of yours using a drone look like?
Now that we have RTK capability with our eBee X, we typically start out by establishing a survey monument in the area of our takeoff and landing site. We then survey that within a centimeter-accuracy GNSS receiver that can provide either real-time corrected or post-processed coordinates. We then assess the area that we need to map by field reconnaissance. Then it’s back to the lab to lay out the flight blocks in eMotion. In the case of our 800-acre project site, we have two large flight blocks, each of which can be covered by our eBee X in a one-hour flight. Once students have drafted a layout of the flight blocks in eMotion, they will typically send it to faculty to see if everybody agrees with the plan.
The workflow is simple. The next step is to go out to the site, lay out some check points and then launch the eBee. On our most recent mission, we flew the 800-acre site with the S.O.D.A. 3D camera and produced a very high-quality orthoimage that we now use as a base map for field data collection using a GIS-based app. We also now have a very good DEM to use for running hydrological analyses.
As someone who works in an academic setting, how has integrating drones into your program helped your students?
For the kind of work that we’re doing right now, it’s another teaching tool. It’s another thing that students learn how to do exclusively in our department. It helps them go out in the world after graduation and go right to work in fields such as precision agriculture, environmental monitoring, ecological restoration and emergency management.
When we began droning up with our senseFly equipment, all necessary training and support was readily available.
Of course we want to foster their fascination with the science of remote sensing, but students need to be able to get work after they graduate. [Using drones] is giving them a marketable skill that they come out of school with. And some of our former students even come back to us after working in the geospatial industry to earn an advanced degree, or maybe they’re interested in teaching some courses. For an instructor, this is really satisfying.
You had a similar experience, correct?
Yes. As previously mentioned, I was an engineering geologist for 31 years and made a late-life career change. I was 50 years old when I came back to Cal State Long Beach and got my master’s degree in geography with a specialty in GIS and remote sensing. Shortly thereafter, the department invited me to teach a few classes, and so here I am. It’s a totally different world, and I’m enjoying it a lot.
When it comes to data collection, how important is accuracy? How accurate was your data collection before using drones and how accurate is it today?
I’m glad you asked that question because I was going to segue into that conversation. Think about this. When you’re just taking images of a location and it’s a one-time thing, you’re most likely fine with just a high-resolution image with a planimetric accuracy of three or five meters. If you want to start taking measurements from the image, you typically want something more accurate. But if you start doing the kind of research that we’re doing, that is, comparing land cover change over time at high spatial resolution, positional accuracy is critical. For example, if we generate an image mosaic of a study site in February and then again in May or June, and we want to compare the difference in the vegetation vigor, we have to make absolutely certain that the images from the different dates are co-registered as closely as possible, pixel for pixel.
Now here’s the problem on this study site I’m talking about. We’re working on a very rugged 800-acre ranch. The terrain is very steep, it’s very rocky, and it can be very hot up there. It’s not a pleasant place to work in the summertime. There are a lot of rattlesnakes there. To send students out and lay ground control points all over this site is totally impossible, but yet we need our images to be positionally accurate.
So, how do you do that? Well, that’s where RTK comes in. We are looking to images from various dates to where they line up eight to ten centimeters, that’s what we really need. And we need to do it without having to spend several days laying out ground control points in hazardous field conditions. RTK is really the thing that makes it possible.
Was it difficult for you to adopt drone technology?
Getting up to speed and learning how to use drone technology was generally pretty easy. Obviously there are always challenges, but when we began droning up with our senseFly equipment, all necessary training and support was readily available so it all went quite well.
Thanks for talking with us today, Scott!