LiDAR Sensors Meet Florida Citrus
The farming business has made great strides in crop management efficiency over the years. Thanks to bright minds and brilliant technology, numerous tools and increased access to information have done a great job of helping agricultural professionals make more informed decisions.
One of the most recent ways proven to increase efficiency in agriculture is through the implementation of precision technologies, often referred to as precision agriculture, which has opened up a huge market for geospatial intelligence. Add the burgeoning small unmanned aircraft system (UAS) to the process and the ability to gather spatial data on large expansions of vegetation has become as efficient and economical as ever.
Precision agriculture remains in the early stages of adoption and, according to Tom McPeek, CEO and founder of AGERpoint Inc., a lot more can be done to spread it into a sub-market of farmers who grow permanent crops that develop on a tree, vine or bush as opposed to annuals that grow in a field, often low to the ground.
Four years ago McPeek founded AGERpoint in Daytona Beach, Fla., to address that sub-market’s needs using his experience with LiDAR. The agricultural technology firm specializes in field phenotyping and permanent crop analytics. So far, business is booming. “This time last year, I employed two people. I currently employ 21. I will add 20 more this year,” he says. “I don’t see going back to teaching any time soon.”
Architecture to Agriculture
Quite often, we start down roads that we never really see coming, McPeek says, and that was certainly the case with his AGERpoint adventure. Rewind at least four years, before the founding of AGERpoint, to a lecture he was giving on the use of LiDAR for building documentation. At the time, he was a professor of architecture at the University of Central Florida and a practicing architect specializing in the use of LiDAR and other sensors to document buildings.
After the presentation, an architect approached him and said, “That’s kind of cool, but could you do the same thing with a tree?” McPeek responded, “For us typically in remote sensing, the trees are the enemy and we’re trying to get to the buildings. That’s not something I do.” But she kept pushing, at which point he said, “OK. I’m a researcher. I’m game. What do you want to do?” She replied, “Well I’d like to introduce you to my husband. He’s a citrus grower.”
A week later, McPeek met her husband, who owned and managed tens of thousands of acres of oranges and grapefruit in Florida. The grower pointed out that if he actually knew how many trees he had and certain attributes of them, he could better predict yield. He told McPeek that permanent crop growers like him didn’t’ have the ability to access a lot of the smart technology that could be found in field crops, probably because the more complex three-dimensional aspect of the trees made his crops more difficult to measure.
“Then [he] took me into this room adjacent to his office and showed me there was a couple of college interns in there and they had these enormous rolls of paper where they had printed satellite imagery of his orchards and they were going in and measuring these trees with a ruler and counting them,” McPeek says. “He said, ‘This is as high tech as it gets in our business and there’s got to be a better way to this.’ So we began a journey, which was to take our technology using LiDAR and go out and document an orchard.”
McPeek did a quick pilot. He took a tripod-based scanner and spent about eight hours walking down a single row of citrus tracking about 80 trees. Scans were taken about every 30 feet. “What we found after we completed that scan was, not only could we get the account of the trees and a basic size of the trees; we could get a lot of things because we could clearly see fruit, we could clearly see the terrain and at that point the light bulb truly came on. Then we were off to the races.”
Fast forward about six months, McPeek quit his position at the university and decided to pursue crop data acquisition full time. Knowing that using the tripod-based scanner wasn’t going to work, he went through the process of filing for patents and developing new technology. He moved into a mobile configuration where he mounted the sensors to the back of a truck. His team went through the process of testing around 25 different types of LiDAR until they found the right technology to do the job.
Fast forward again four years; now McPeek’s team can scan an orchard in an afternoon. Typically they can cover up to 300 acres a day from the ground and about 3,000 acres from UAS. The different approaches result in different levels of granularity, but they are able to collect a lot of information either way.
McPeek calls it a field-based phenotyping solution. “We get the morphology, which is physical attributes; the location of the tree; the diameter of the trunk; height of the tree; diameter of the canopy; and density of the canopy. Then we add phenotypical information, so we can tell you what variety that tree is and whether there’s presence of stressors. Those stressors could be disease, they could be insect infestation, they could be drought. If it’s disease, we can tell you which diseases they are.” This year, he says AGERpoint will introduce technology that allows them to find and count individual fruit on the tree.
Most of the crops AGERpoint deals with are planted in rows, whether it is an orange tree or a blueberry bush or a grape vine. Most locations can be navigated between those rows. As a general rule, they like to scan from both sides of a plant because it gives them more information. “Typically I’ll say that if we go on the ground, take the vehicle through the individual rows, we can get to the granularity of a seed. If we go from the air, typically we’re going to get to the granularity of a tree.”
‘LiDAR is the Core’
McPeek says that there is a lot of technology to choose from and that you get what you pay for. The LiDAR sensors they use are generally RIEGL. For aerial capture, they have a 333 exemption to fly a 3.3-meter fixed-wing Penguin B drone. It is gas-powered and can run for about four-and-a-half hours in a single flight, carrying the exact same sensors used on the ground.
To present data to clients, the firm uses a full SaaS-based platform that rides on Esri infrastructure and includes a full GIS-integrated mapping solution. Since a lot of the clientele are not GIS specialists, they create a lot of custom algorithms to make the data easier to understand. AGERpoint does both 2D and 3D shape files and the data processing stage is 90 percent automated, McPeek says. Most of their hardware is commercial off-the-shelf, while the software is all proprietary.
“Ultimately that’s the foundation of what we do. LiDAR is the core of how we do what we do, and it is obviously my background. I’ve been doing it for 15 years and it allows us to get to a level of granularity that is unmatched by anything else,” McPeek says. “The current state of the art for growers is either NDVI or multispectral imaging from aerial photography. That gets you a granularity that may go sub-foot —typically does not — but its strongest limitation is shadowing. The beauty of LiDAR, depending on which technology you select in LiDAR, is it’s not encumbered by shadowing at all, so you can see things that you shouldn’t be able to see.”
On the day McPeek went to meet the Florida orange and grapefruit farmer — the start of the AGERpoint journey —the grower told him that there was no reason he shouldn’t be able to walk up to one of his trees that had been there three or five or 10 or 100 years, locate the asset by GPS alone and learn everything about it — when it was planted, when it was treated, what was harvested and what was applied to it.
“Every product that we’ve developed up to this point has been at the request of a grower saying, you know it’d be really cool if we could do that.” As of today, AGERpoint has introduced everything that grower asked for, which has manifested into a tool that allows the grower to take their phone, stand in front of their tree and take in all of its characteristics.
McPeek says the tool isn’t limited to offering growers information access; it allows them to input information too. “If I’m in the field as a field scout and I see there’s a problem with that tree, I can digitally flag that tree and that will send a tasking order through the database and allow us to do labor tasking based on that initiated response. So it’s how do we take and do full-circle data analytics for our consumers, which in most cases is growers.”
To highlight just how important the solution is to his clientele, McPeek says it’s important to note how off any given grower can be when estimating yield without the help of LiDAR. He says a story his first grower told him was that he’d been driving around the field all day with his grove manager—they’d been doing it around 30 years — and they finally stopped in front of a tree. They had been talking about what the harvest looked like and what they thought it would ultimately yield, so he turned to his grove manager and he said, “OK, that tree right there; how many oranges are on that tree?” The grove manager said there were 450 on the tree, which McPeek considers a really good, safe guess. The grower replied, “That’s fantastic. Get the crew over here. Let’s pick it. Let’s see what’s on that tree.” When they finished picking it, they counted a whopping 1,450 oranges.
“This is important because this is a juice orchard. Juice orchards don’t sell it by the bin; they sell it by the field,” McPeek says. “So had they gone based on that single prediction, they would have given away two thirds of their crop that day. So understating what truly you have in the field is a powerful thing for growers. It’s a powerful thing for consumers and ultimately that’s what we’re looking to help them understand.”
Service to Software
While growers have been the main market of AGERpoint’s LiDAR-powered solutions, the firm serves a wide range of clientele. They offer tools that help illustrate food traceability for consumers who are becoming increasingly interested in where their food came from. For bankers, real estate agents and insurance companies focused on best understanding asset value and makeup with regard to properties thick with permanent crops, the data laser scanning provides is invaluable. Furthermore, the solution is relevant to the commodities market because the data set can help predict what yield will be in a traded commodity.
With so much client potential in so many different markets, McPeek says he has learned over time that the tools AGERpoint is developing are capable of covering a much broader range than they can as a service provider — what they initially set out to be. For that reason, they’ve begun developing a network of partners, many of them in the geospatial community, interested in actually collecting the data. “Then we can step in and say, we’ve developed all of these tools and algorithms and all of these patented processes that we can help you in delivering that data to your customers now. ... We still do collection in the field, but we’re beginning to do less as other people begin to feed into that stream.”
LiDAR: The Next Industrial Revolution?
As far as McPeek can tell, the future looks bright. AGERpoint is involved with projects in a variety of countries and he expects continued growth within each of those marketplaces. He can see the firm doubling in size every year for the next three years at least; and while his career and AGERpoint’s approach have changed a lot over the years, he says the relevance and value of LiDAR remains.
McPeek says he recalls reading an article a year or two ago about LiDAR being the next industrial revolution, and some saying that may have been overstated. “I’m not sure that it was strongly stated enough,” he says. “There are enormous opportunities that come with the power of LiDAR, the ability to quantify and measure almost anything, and I think that the space that we’re working in is a perfect example of that. If somebody would have told me 10 years ago that you could drive by an orange tree and in the mere seconds it takes to drive by it, you could tell what variety that tree was, what the health of that tree was, how many fruit were on that tree, and what the predicted productivity was going to be over a number of years, I would’ve said you’re crazy. But that’s the reality that we live in.”