 I don't hear the boarding music, Jerry. Okay. Okay. Friday afternoon, folks. Ted Rawlson here in our downtown Honolulu studio, actually overlooking the Kauai Coffee Company by remote here on our show where the drone leads. And today our drone show takes us right to the freshest cup of coffee you could possibly enjoy here in Hawaii out of Kauai Coffee Company. Our guest today, a longtime participant on Think Tech Hawaii when he was in Honolulu now on the island of Kauai is Fred Cowell, General Manager of Kauai Coffee Company. Welcome aboard, Fred. Thank you, Ted. Welcome. Good to have you here. And we just met this week at the Capitol. It's been a very busy week here in Hawaii regarding drone activities, hearings in the Capitol, the aviation day in the Capitol, ag day in the Capitol. And yesterday we had an event down at UH where we had demonstrations and displays and discussions on drone technology for members of the legislature. So anyway, but Fred comes to us with a really exciting perspective and use case for drones in a very positive way on the island of Kauai in terms of optimizing crop progress in the coffee plantation business. Fred, tell us a little bit about that. Well, I've been with Kauai Coffee for about a year and a half. I came from oceanic laboratories doing a bunch of infrared research. I knew that through sensor technology we could help our farm operations dramatically. We've got roughly 3,100 acres and we cover it with manpower. And the ability to get up in the air, see things that we can't see from the ground, monitor things that we can't monitor from the ground just seem like a perfect fit. So you've been able to get this started and begin operations with drones actually today in Kauai. Well, we began using them for overhead imagery, trying to create orthomosaics. Using near-infrared, we wanted to look at plant health. We wanted to look at irrigation stress. We even wanted to get the spot invasive vines from the air. We found out very quickly that it works well not only for aerial mapping but also for video or even photospheres, which are very handy for virtual tourism. So from the perspective of managing the crops and managing the water, managing fertilizer, looking at the basis species, and optimizing the production of the coffee product, you found this just in a very short time making a big difference. We do. Even from the very first photos we took, we were able to see specifically where irrigation lines had failed. We have one of the largest drip irrigated coffee farms in the world, roughly 2,400 miles of drip tape emitting irrigation and fertilization through the drip tape. If one of those lines is cut, eaten by a pig, damaged in some way, we could have hundreds if not thousands of trees that would go short on water and nutrition. Unfortunately, we don't see that water stress or nutritional stress from the ground very well, and we want to be able to spot it before it becomes critical. So that's where overhead imagery, specifically multi-spectral infrared industry, gives us a whole lot of additional tools. This is pretty interesting. Just a couple of years ago, this technology wouldn't have been available. That is the software that does the processing and does the various photo mosaicing and various multi-spectral imaging and the composition of the expressive layer would not even have been available except for a very high-end system. So what's happening is the good old Moore's Law is favoring us now. Things are coming down in price, up in performance, down in weight, and you kind of wonder what's going to happen in the next two years, Fred, as this all moves forward. Well, from watching this from three years ago, I recognized very quickly there were three things that were working in our favor. We knew that GPS technology was getting us tighter and tighter control of the drones even down to the millimeter level. We knew that camera technology was improving in leaps and bounds, and we also knew that the software to control both the cameras and the aerial devices themselves was getting developed much more rapidly and is much more usable. For instance, on my drone, the controlling software that I'm using just for the flight control software was a $20 app. Some of the image processing software is done by subscription, so it's not something that I have to have in-house, and all of this is now becoming commercial off-the-shelf. No longer experimental. So you can actually turn this into almost a farm as a system. It's almost the orientation you're generating here. In fact, you're seeing patterns that you never saw before or that the farm never saw before. So we don't even know what's going to develop in the next couple of years in terms of how these patterns can turn into positive indicators and just thinking differently about the whole agricultural situation. Right. Another way to think about it is we are, although there is agricultural land available, not all of it is well-suited for growing coffee. People say, well, if you have increased demand, can you increase your acreage by 10, 20, 30 percent? That's an extra impossible. I don't have the infrastructure. Land is expensive. Irrigation infrastructure is expensive. But it is possible through some of these technologies to see increases in yield by 10, 20, or 30 percent or decreases in production costs by 10, 20, or 30 percent. It all adds up to the same number in my opinion. I would rather do better with the information and the land and the labor force I have than try to expand into greater, greater acreage. So capitalize on what you got available today. Get as much out of that as possible in a balanced way and use the technology to help you get there. So at your scale of 3,000 acres, this works out very well. How many drones have you got in the fleet? Say that one more time, Ted, I'm sorry. How many drones you have in your operating fleet of drones? Right now, we've just got a single drone. So what? When we went out to purchase this drone, we set a specific requirement. Our specific requirement was based on number of acres per day to map. And our goal was to be able to get 100 acres per work day. That would give us the opportunity to do the entire orchard in roughly every two months, assuming that I'm fulfilling every other day. Right now, my most current imagery that I'm using for farm management, my GIS, is three years old through Google Earth. I've got the ability to now put updated imagery into my GIS wherever I need it most. So you can take your existing legacy GIS picture and drop in what you've done yesterday and the day before and continue updating that mosaic with current information. Correct. Or you can layer it over time. So you can look for trends as well as for patterns. This is a complete different way of looking at farming, isn't it? From what people have thought of before? It is. Let me give you an expression that my father taught me once. He said farming is easy. It's just a matter of doing the right thing at the right time. And what the drone enables us to do is shorten that information loop so that we can do things at the right time. That's all we're doing. The circle of information, the ability to get information back into the decision loop is really what we're trying to shorten. This is a good old-fashioned Oodle. We all learn about it. OK. And so at your scale of 3,000 acres and 30 days to process the entire, recycle the entire farm imagery, you're beginning to see a pattern of operation that's going to be useful to you. If you had to go with greater coverage, you need more drones. If you had a reliability situation on the drones, you need to spare. There's other things that will come out of this experience that will be useful to all of us. That's exactly right. The other thing that we're recognizing, and you alluded to it earlier, there are things that we don't even know that our drones can do yet. We've been having some brainstorming sessions. We've got horrible problems with pigs coming in and tearing up our fields. Pigs with a thermal sensor are easy to find and direct wrappers to take care of them. Otherwise, they're just out there somewhere. And they're smart enough to know that when the drip tape is full, they can bite it and get a drink of water. It's like a drinking faucet for them. The other thing that we found very quickly is people are just absolutely interested, not so much in the name of the coffee or even where it came from. They want to know what you're doing. So we can now give information about what's going on in our fields near real time so that our customers feel part of what we're doing. They're not just out there in the store somewhere. They're on our farm. You could put virtually real-time information dynamic on your website. And people could see that very coffee they're going to drink two weeks from now sitting in the process somewhere. That's pretty, this is amazing. And if you think of, again, your scale is 3,000 acres and with one piece of equipment, that's pretty darn effective in terms of being able to get one several thousand dollar drone to do all that work to improve your productivity. If you were to think of scaling up to a much larger farm, or if you could think of scaling down to the two acre truck farm, how do you see the drone activity fitting in those pictures? Well, there's always the right tool for the right job. In our case, we were looking at a specific set of requirements and we sized our purchase based on that. There are licensing requirements. I went and got my FAA remote pilot license in Honolulu a couple of weeks ago. Certainly worthwhile, highly recommended, in fact required for anybody that's using it for commercial operation. So as I spend more time in the front office, my orchard team is gonna become licensed so that they can fly when we need to. It's also a matter of opportunity. When the sun is right, when the wind is low, when the conditions are ideal, we gotta go fly. Multiple drones is perhaps the next step. We could go to a fixed wing drone, which would cover more acreage. And there's even folks that are using helicopters and man fixed wing to do the same type of survey. We just don't quite have the flexibility. On small farms, which couldn't afford to buy perhaps a multi-thousand dollar drone, there are folks that offer farm survey services on an hourly basis, a great way to go. And you don't have to buy the equipment and you get somebody's experience of what they do. And then what you would need to think about is of course, as the scale comes down to the two acre type farms, the imagery collected would actually pick up somebody else's farm as well. So there's issues of exclusion of information. Certainly wouldn't wanna have your farm being looked at by somebody else's drone. How do you see that all coming together in terms of confidentiality, preparatory information, that sort of thing in terms of this kind of collection? That's a little bit tougher. Fortunately, coffee trees don't carry that many secrets. Provided your neighbor is doing everything that's completely above board. If I were on a smaller farm and I didn't have a neighbor that was exactly next to me, I would let him know that I'm flying on a specific day, just out of courtesy. Just common sense wouldn't necessarily need to prevail. The other thing about a small farm with two acres, rough number of trees on two acres is maybe 1200 to 1500. I could walk my entire farm in a couple of hours and get the same information that I put from the drone or similar type of information. I just don't have that many legs. Again, 2400 miles of coffee trees, that's a lot of walking if I were to try to look at. So the smaller farms would want to bind together and hire a service from some service provider. So they wouldn't have to spend those two or three hours walking their two acres, but the drone would do it in five minutes for that small acreage and then just combine it and then separate results out of the processor and provide it to you as a service. Right now, my survey with Full Earthal Mosaic is approximately 15 acres in 15 minutes. So an acre a minute. That's a pretty darn high rate of accomplishment that seems to me, an acre a minute. That's, you can't even, I mean, in a minute, you can't barely penetrate that acre of walking. So the- Yep, that's a bit tough. So if I were walking, it would be, my conversion is roughly 4,800 feet for every acre of trees. So roughly a mile walking for every acre that I'm doing. So you have the speed. The thing about the new technology is we don't fly it. The new technology allows us to program the data gather, send it on its mission, we're required to be in control, have it in visual range. But it flies its mission and then comes back when it's complete, we'll swap batteries and go do the next one. Let's come back to the, to how this might extend again into the education domain and to the community awareness and this sort of thing when we come back from our first break, Fred. Well, one of the biggest issues that- Hi, and thanks for watching Think Tech Hawaii. My name is Justine Espiritu and I host the Hawaii Food and Farmer series with my co-host, Matthew Johnson of Awaku Fresh. Every week we bring on farmers as well as all the other individuals and organizations that help support a thriving sustainable food system. In fact, it's interesting to learn what others are doing so you don't have to be a Hawaii resident or producing food on Hawaii to be featured on the show like today's guest, Wyatt Bryson of Jewels of the Forest and Michael Lab Solutions. Aloha, thank you. It's been a pleasure being on the show. I love seeing what you guys do and I really support your mission and it's really nice being back in Hawaii and thank you again, it's an honor. So you can see guests like Wyatt every Thursday at 4 p.m. on Think Tech Hawaii. Thank you. We're back on live folks, second half of our show here where the drone leads, Ted Rawlston and Honolulu. We have Fred Cowell from the general manager of Kauai Coffee Company and Kauai in his office. Fred, welcome aboard again. Thank you. Okay, and by the magic of Skype, which is working now, we're in good shape. We had some troubles earlier and that's another issue that comes up. We have complex electronic systems generate interface issues and problems that we have. So your experience will be great because it'll be showing how these complex systems executed in the form of a drone and a ground controller work over a long period of time over different weather conditions. You're almost part of the pan-Pacific unmanned air systems test range, Fred, by the work you're doing. And we actually need to tap into your experience and use that to drive some of our tasks here at the test range. But what we were speaking of just before the break was the nature of scaling, the nature of how a single drone is so effective at the work you're doing and how that effectiveness would be translated to larger and smaller farms. But the other piece we want to talk about translating to is the future of farming and the interest that our kids have growing up and how educational programs in the schools could learn from what you're doing and help take the next step. So how do we increase the interest in farming in terms of the, not just the farming itself but the increasing production of farms through this new technology? Well, farming is by its very nature becoming more and more technological. Much of what we're doing, we believe we can integrate sensors so that most of your meters and emometers for wind rates, solar measurement, tell us what's going on with the sun on any given day. And all of that requires some infrastructure and talent. And as we begin talking with our school age kids, very difficult for me to get kids interested in getting their hands dirty but for our firm a chance to go fly a drone or work with a computer then they'd work out. So that's cool. So the drone has to appeal in itself just by its very nature as well as the value it provides. So turning this into student programs, tasks and such that the high schoolers can get involved in, taking that integration you're speaking of which just adds to the system definition of the system complexity and completeness of the farm as a system. So you're kind of a situation where the information is being generated maybe even faster than the system orientation can accept it. So we do need some high level problem solving thinking at that level of integration as well as the sensor development and such. Now this is a almost becoming a complex adaptive system problem which all of us love to take on. It's, we can even say that's a system of systems. I've got an irrigation system, a fertilizing system, a nutrition system, harvest system, processing system, all of those things come into play. None of them stand completely alone especially in a company like ours it's vertically integrated. So as we take that system of systems approach the exchange of information really becomes key. And this is again something that without the, oh here we see some graphics of what we got here. So we'll talk through this and can you see it Fred on the screen? Okay this is the chart we used this yesterday at UH. We're showing here the use of UAS and Agriculture and Coli with your company logo on the middle. Showing a couple of distant view graphics which I think as you suggested would be very useful in communicating with customers and with suppliers and such on just what the farm looks like today. But then we show on the lower left the green leaf index which I think you flew last weekend and tell us about the green leaf index analysis Fred and what the red bands on there mean to you. Actually I shot the green leaf index yesterday at about nine o'clock. Okay. And was able to turn it and give it to you by about noon. Even with processing in the cloud. What we're looking for in that particular block we knew we had a vine infestation, a monolow vine that was choked in some of our coffee trees. And we were doing some experimentation where we could see the thickness of the vines or the difference in the vegetation based on the vines. The green leaf index is a little bit simplistic for the challenge. It's showing us basically where I've got the greatest amount of leaf mass. Where I've got the red showing through some of the lines that's thinner foliage and where I've got the darker green that's the dark foliage. Or excuse me, the heavy vines and leaf foliage. So what we have here though is the beginning of a pattern identification of what's going on in the ground, in the watering system and maybe even in the atmosphere and certainly in any invasive challenges that are here. And so if algorithms could be brought forth that would give you sort of instruction as to what to do with that once we have the history understood and the ground truth and what we're gonna do about it. So that basic presentation of data and that first level of filtering is a key to what the next step has to be. And you know- That's correct. One disadvantage that I've got is there are many coffee farms in the world. I don't know of any that are doing coffee leaf algorithms. I've got a lot to learn from the wine industry and other industries in California that are of similar scale. But I'm kind of out here on my own and my next shot would be to take that same field with a new infrared or some other infrared spectrums and see if we could break something else out and then tie it back to the ground truth of what's going on in the field. And so what that's all about more and more is that the drone is a skyhook holding a sensor and what's critical is the software that does the analysis and does the expression of what the information is to let people make decisions. That's why the logo or the label and the statement in the middle of that chart was hey it's all about pixels but it's more about what you do with the pixels. I mean a picture or even a video is nothing more than a gigantic pile of pixels and how we filter them, how we take out the various reflectance aspects and combine them with other pieces of information to try to get some inferences here. Things that our brains can't even do can be done now operating on the pixel package. So I think a lot of our future is in getting our kids to think about that. What kind of filtration, what kind of analysis, what kind of combinations are necessary to get the best picture out of that. And anybody involved in the dynamics would know that the common filtering domain which is how the same sort of thing is applied to dynamic streams of information has so many different filter packages that are available today and they're changing every day. So it's all about taking that information in the pixel pile and doing something with it. And you know we don't think about that much. Well I think from now on I'm going to explain that picture of green leaf index as a brain scan of my coffee order. Yeah, okay. And then let's talk about the picture on the right in that chart that we showed a minute ago and that's the landform analysis which is incredible. From a plain old camera we're getting a landform analysis which talk us through that red that's the volume estimation piece where we're looking at cuts in the earth here. Correct. One of the piles that's highlighted there we're trying to estimate the volume of that pile. We have an on-form composting operation here on Kauai Coffee that does almost five million pounds a year of waste pulp. That waste pulp is extremely valuable in our orchard to spread micronutrients, microbiology and increase our organic matter. The ability to estimate the volume down to the cubic yard is a huge help. Now I've got an ability to determine how much compost I've got ready or how many acres can be treated based on the size of that smart size of that pile. That's done through a series of shots measuring ray tracing from various different angles to come up with a composite volume. They say that the elevation bands or the ability to determine the elevation of that pile is to within centimeters. So that pile that circled, we came up with 2477 cubic yards based on that flight. And so that sort of thing would be extremely useful beyond farming in the construction industry, in the quarry industry, in any kind of land mass looking. Even things like disaster management and disaster operations where you want to predict like in Orville, California right now where landslides are an issue with the sod and earth. So looking at landforms, determining the forces on the land and this sort of thing, all these things what we have to ask our kids to think about and how we want to take that pile of pixels and do something with it to generate the kind of insight and information you're getting there. But this has got to blow you away, Fred, to what you've done here in just a couple of days. I mean, it should blow anybody away. And at the commercial price that we're talking about these days, this information is just never, we've never seen this before. It's the speed of change, that's really what it is. Yeah. And Moore's Law will say that this will be available free on the internet in two years and then in the next wave of sensors and sensor analysis will be coming into the picture. Is it possible, do you think, after the short experience you've had here, projecting forward, could you write down what the most desired output would be of this sort of pattern analysis or OODA loop optimization? That would be it. Sorry about that, but... I almost believe I'm frightened to speculate, but I'm absolutely curious. There are, and it's not gonna come from me, it's gonna come from somebody else looking at the problem from a different angle or who has seen something similar used in another context that says, hey, let's try that. There really is no limit, it's just a matter of how creative we're willing to be. I run a poor profit company, I have to maintain budgets and timelines and production lines, yet, if I can leverage technology to help me get to those goals quicker, I certainly wanna do it. You know, we've worked before, and we've had on the show here, some of your colleagues from your prior company, Oceanit, and there's some things coming in the way of image extraction or information extraction from video and from stills that would even take the looking at the picture out of the game because there's no reason that the filtering systems can't recognize things like those leaf index lines you were looking at. And so it takes the human out of the loop, you can still go back and review it, but you get a recommendation by the automatic analysis. And at UH, one of the student programs is even putting that on board. So we don't even drop the whole video signal down to the ground, just do the analysis in flight, on board in real time and produce a thumb drive at the end of the day with the answers in it. That's a search and rescue activity, but the hardest thing we've all got is the question I asked you, how do we figure out what the requirements are that we need to write down for the benefits of the future and get people pushing in those directions? And the same applies and we're dealing with law enforcement, we're dealing with at the federal level and the state level and the environmental issues, coastal geology and such. And so it's gonna take some experience and some leadership and this drone rotary that we're talking about that we're gonna do in your property in October to try some of those experiments and tease up the issues of what the future might benefit from. This is a pretty exciting time for a guy like me to be seeing it all happen. Well it certainly is, I began my Air Force career shooting Celestial out of the top of a C-130 to find my way across the ocean and now GPS can define me within a quarter of a centimeter. Things change, things change for the better, but it is certainly an exciting time. The old expression that whiskey is for drinking and water for fighting over means that the ability to use drones and other tools to use water more wisely, huge area of interest. That's fantastic and certainly in Hawaii where we are transitioning from plantations and some areas to golf courses, so that's that whole issue of water use is a big issue. Over and over again we see that. Anyway Fred, thanks so much for coming on board and thanks to the leadership you're generating here. Without this kind of real activity that the benefits wouldn't be evident and they'd be talking. You're putting them into real terms and that is such a positive leadership step here for the whole state, so thank you very much for that and we'll get you all back on the show. Once a month or something like that, I'm sure you've been making that much progress that we'll learn so much from you, but thanks for coming on and we'll see you again at the Capitol I'm sure, very soon and certainly at the drone rodeo on Kauai coming up on October. Well I hope I'm able to host it for you, we've got a wonderful site. Okay that's great, we're looking forward to it. So folks, that's it for our show for today and Fred Cowell, thanks for coming on and we'll see you all next week.