When it comes to planning drainage activities on farms that grow crops such as rice, elevation survey data—and its accuracy—is a big deal. However, using ground-based surveying instruments to create the digital elevation models required, in a timely manner, can sometimes be a challenge. That’s why Ryan Moore, the crop services manager at Greenway Equipment, Inc., embarked on a new elevation survey study at Senter Farms in Osceola, Arkansas. His aim: to assess how RTK-enabled survey drone (or UAV) technology might replace, or complement, the use of ground-based GPS devices when mapping the terrain of a rice farm.
Prior to the official publication of Greenway’s full report, Waypoint caught up with Ryan and his client Travis Senter, of Senter Farms, for an exclusive sneak peek into this unique project.
Hi guys. Let’s start by learning a little more about Greenway Equipment Inc. What exactly does the company do?
Ryan: Sure. Our job at Greenway is to find solutions for customers that help them to be more productive and efficient. We take great pride in finding them the right solution. These solutions can be anything from hardware to software, to helping them through our data management and data analytics that we do in-house.
Could you give us an overview of the study you ran? How did this comparison of elevation survey technologies come about?
Ryan: Sure. We work with Travis, the assistant farm manager at Senter Farms in Osceola, Arkansas. Travis is growing lots of rice, so there are lots of levees that have to be built. And while over 90% of the ground is irrigated, there’s work to be done to irrigate that more efficiently.
… the flatter your site, the more accurate elevation you need
Of course, the flatter your field, the more accurate elevation you need—you can be more forgiving on a large elevation change—and Travis’ field has a pretty flat elevation. So, we need to work off a highly-detailed and very accurate elevation survey.
Traditionally, would you employ terrestrial surveying instruments to collect such elevation survey data?
Ryan: That’s right, yes. We would typically use RTK GPS equipment, with us getting into the field on a vehicle of some kind, like a tractor, a pickup or a UTV [utility terrain vehicle].
For us, it was interesting to conduct this UAV versus ground RTK study in winter, because in Arkansas we have lots of rain. This means lots of downtime, when we could be surveying fields—for designing levees for example—but it’s too wet to get into the fields and get this work done. Obviously, we’d prefer not to have to cram that work into the early season, in the springtime.
… the idea of trying UAVs was that these aircraft might be useable when it’s otherwise difficult to get into the field on the ground
So, the idea of trialing UAVs was that these aircraft might be usable when it’s otherwise difficult to get into the field, on the ground. It would allow us to get ahead of the game and be more efficient, meaning less wear and tear on our other equipment as a result. However, we needed to know that the quality and accuracy of the UAV data would be good enough to make accurate water management decisions and designs, specifically earthwork, rice levee, and field drainage ditch placement designs.
So, could you tell us about what you looked at, specifically, in your project?
Ryan: We ran two elevation surveys on the terrain at Travis’ rice operation: one with a typical ground RTK GPS, and one with a UAV.
We wanted to see how closely the elevation collection methods were to each other on a 40-acre rice field that was recently graded on a single-plane slope
To collect our aerial UAV survey data, we borrowed an eBee Plus, with activated RTK/PPK, from senseFly (thanks, Adam and Troy!). Then, Travis allowed us to use his Trimble GPS equipment; in fact, he collected all our ground survey data for this study.
We wanted to see how closely the elevation collection methods were to each other on a 40-acre rice field that was recently graded on a single-plane slope. The field was graded at a 0.1 slope at a 0° heading.
We surveyed the field first with the senseFly drone, carrying a senseFly S.O.D.A. RGB camera. We used a Topcon Hiper SR GPS receiver as the RTK base for this study and the flight took approximately 20 minutes, flying at 400 ft above ground level with an 80% image overlap.
The field survey . . . had to be done two days later due to the wet field conditions
We then loaded the images into Pix4D software for processing, exporting a digital surface model (DSM) point cloud format that our land forming design software, PTS T3RRA Design, could use. The processing of the flight’s images took approximately three hours.
The field survey actually had to be done two days later due to the wet field conditions at the time of our UAV survey. The ground collection was carried out with a Trimble TMX-2050 display and Trimble RTK GPS receiver, to give us a comparison for our current surveying practice. The survey collection pattern was done on a 50-foot distance (running north and south) then roughly 100 feet on an east and west pattern. This survey took just over an hour to complete.
And are you able to give us a sneak peek of the results, even though the study is currently being prepared for publication?
Ryan: Yes, we can give a summary overview. On one hand, you have the obvious difference in point density between the two approaches. The GPS ground survey consisted of 4,968 collected points. In contrast, from the document produced in Pix4D at the time of processing, the UAV survey’s point cloud produced 28 million 3D densified points.
… the UAV survey’s point cloud produced 28 million 3D densified points
Ground surveys will then have gaps between collection passes, which leave it up to software to interpolate. These areas leave greater room for error as they are not known elevation values, only interpolated. With a UAV survey, we are receiving consistent readings with, relatively speaking, no gaps for interpolation.
In terms of the positional accuracy achieved, are you able to give us any idea of how the terrestrial and drone data compared?
Ryan: I’d put it like this: there’s enough correlation between the two methods to conclude that the eBee Plus, with its RTK/PPK function, provides a viable source of data for elevation surveys to be used for levee line placement and earthwork designs. There were negligible differences between the two survey methods.
There were negligible differences between the two survey methods
Travis: The results were quite impressive to me, especially considering the conditions on the day of the flight when it was quite windy and cold. My initial thoughts were that this small foam plane, without any sort of gimbal, couldn’t possibly be accurate enough to get the results needed. After seeing the final results though, I was impressed with the level of detail it generated, which was very representative of the ground-based survey.
My initial thoughts were that this small foam plane, without any sort of gimbal, couldn’t possibly be accurate enough to get the results needed…
Drone surveys are, for me, something that will probably change the way we approach land grading. I’m not here to say a UAV will totally replace a ground elevation survey, but it will give extra insights into things you can’t reach with a ground-based vehicle—places like large ditches, mounds of dirt, and even low-lying bayous can now be given an actual yardage for the dirt needed to fill them or remove them, instead of guessing.
And finally, are either of you able to extrapolate what your study’s findings might mean for farmers more generally, in particular, rice farmers?
Travis: I’d say that rice farmers, who need to create levees for each field and terrace the water, can use drones to get an extremely accurate view of their field, with a level of detail that is superior to a ground-based survey. There are certainly times when you have trouble watering certain fields and being able to see elevation on a tight scale is imperative. The current methods we use now have negatives, such as using a survey machine that doesn’t cover every spot in a field, and these machines also have tires, which can cause depressions in the soil and skew the results. Aerial-based machines, with large image overlaps, can achieve better results.
Farmers … can, with drones, get an extremely accurate view of their field, with a level of detail that is superior to a ground-based survey
The only drawback at the moment is the processing time needed to turn the imagery into a data point file. As technology improves I see this no longer being a factor, but at present it is a hurdle to overcome. Generally though, I’m extremely excited about UAV technology and I can’t wait to see its potential within my farming and land grading business.
Excellent. Thanks for the overview, guys, and well done on the study. We can’t wait to read the full report.
Both: You’re welcome.