Harvesting research in automating more of specialty crop farming process

Hi everybody. I'm Paul Yeager. This is the MToM podcast, a production of Iowa PBS and the Market to Market TV show. We're going to be talking about automation when it comes to agriculture machines doing work labor specifically in fields. Now we've heard about auto steer and things like that. But this is a little different. This is automated in areas that are usually pretty labor intensive in the state of Michigan, by the way, they have 300 specialty crops. It takes a lot of labor. And as that situation changes, this all comes. This idea has been long in the works before anything political has come around. This has just been a fact of life many people have dealt with is labor, specifically specialty crops. And we're really going to get into asparagus and how a work in research has helped change the industry or could change the industry if things continue. In the research, we're going to talk with Doctor Yuzhen Lu, who is an assistant professor at Michigan State University, and he has been working on for years the technology to make automation come to the field. We'll talk a lot about asparagus and other things with the industry, but we're going to talk about labor and time and all the things that matter in agriculture or bio research. We also are going to talk about the TV show. At the end of the discussion, we will bring Julian. We're going to have a fun conversation about one of the behind the scenes portions of market to market. So that's what we have in store in this installment of the MToM podcast. You can't compare all land grant institutions, but North Carolina, Mississippi and Michigan. Obviously, Michigan's the coldest, but which one has been the funnest for you so far?

[Lu] I cannot do the direct comparison. I actually enjoy living in both North Carolina and Michigan. I live longer in Michigan. I, finished my PhD in Michigan State.

[Yeager] But all three are different in a way that they view agriculture. And to me, as a for for to me, for you as a researcher would be interesting to help give the greater picture. I mean, North Carolina has a lot of row crops, agriculture hogs, Mississippi row crops, but a whole lot of different types of crops, than than what we have. And then you have Michigan. What is the what? How would you describe Michigan agriculture to an outsider?

[Lu] You know, there's a difference here. Michigan has a very diverse health. I post the cherry also. I like blueberries. So we also have the vegetables. Asparagus you know, is that is yeah. Very diverse. There's many options for me. I like engineer, I can work who is not just focusing on like it, just like in the Mississippi only cutting soybeans. That is, is all spread across the challenges is more actually so at a higher level.

[Yeager] With the specialty crop, it comes a major challenge that in California, Florida, they have found they are very labor intensive, intensive. Whereas in corn and being country we hop on a tractor, we hop on a combine. There's a sprayer involved. I grew up walking beans and picking up rock. It just doesn't happen as much as it used to. What are the challenges of the specialty crop from an agricultural standpoint?

[Lu] Yeah, you know, there are the labor challenges. That's the number one challenge for most. I want to say it's very new crop growers. They are not able to sustain without technology to help with the labor costs keep increasing. So, you know, one major production cost very big a portion is the harvesting. Compared to many other operations on farm. Harvesting can cost, over like 50, even, 60% of total production cost. This is a very big, for growers, very big challenges to deal with.

[Yeager] And it's not getting any easier to find labor. It becomes labor itself becomes a political issue. But labor is a challenge that I mean, even in corn and being country, there's not the same number of people to work on the farm that there used to be, and in other areas there isn't. So how does technology help, solve an issue that a farmer might have.

[Lu] To get technology in if they can help? well, on one hand, reduce the dependance on labor. Some tasks can be automated. that were is almost still, take the harvesting as example. If we can reduce the labor dependance, have some automation or mechanical system. may not be fully replace all the manual harvesting. can supplemental be very important or two in case, the goods cannot have enough labor supply. They have a technology available. They can use. Otherwise they have to fully rely on the many workers, many workers now available. They are not able to harvest. They are the total loss of their youth.

[Yeager] Well, go back to Covid 2020 era when the crops, the labor wasn't there because they they weren't allowed to work and then they got an exemption. They were allowed to work. But then it became the routing issue was there wasn't a distribution for them because of whatever issue we could have labor. It happens. It's not just about a, and the actual person to do that, the issue, it can happen where a crop, there's only a small window that it can be picked. And that also creates you are the problem for us. Is technology also help? I don't know if a technology can extend the window or does it just hyperfocus? The importance, of when the work needs to be done.

[Lu] And that window probably by for technology that's the that's one that relate to the breeding or other like horticultural scientist. They if they can develop, you know, variety allow it's a longer window for harvesting. That'd be great. But I don't see that as a viable option. Technology can improve on this. The efficiency, can do the work in shorter time. See, if you have a robot, you can let it run 24/7. So what? Human workers see, you cannot do that. So you can double. If you double, triple the efficient harvesting efficiency, you can reduce the time needed for harvesting.

[Yeager] We'll get into the money savings in a minute, but I want to go to, let's example of asparagus, on market to market. We had a few years ago some video shot in California of an asparagus field and about 12 to 24 people going through this large field, and they have a knapsack on. And it's a it's a long razor, basically that they are spade going through and picking. Tell me how technology can replace that human that's doing that job.

[Lu] Yeah. This is still is I will say so the area of research, but there's, several companies currently. So working on the Harvester, I would still consider the Harvester prototype, but they do show some promise using the vision and the I. So on the, mobile system with the harvest mechanism so the system can localize the spears. That's it. Ready for harvesting and trigger this. The hardware to take the cutting all the other operation to collect the spears. Yeah. This is a, can be done technically by the robots, but it still has a lot of challenges.

[Yeager] How big of a I mean, is this a robot that's on a tractor? Is this something that goes two rows at a time? What does it look.

[Lu] Like so today? So the robots, most of them are autonomous, fully assembled for about, but, yeah, for asparagus harvesting. Actually, the research has been done for many, many years back to even to the 1970s. That's the old. And there's a mechanical harvesting system. Some is the tractor port. Some ways they also can have the human to operate. So for those like mechanical harvesters without reason, there has a very big issue with the yield loss. They normally cut all the spears, cut all the spears. So you may lose 15% of your crop. Because this asparagus grows in the not uniformly. There's, you have to, to harvest, you know, selectively to maximize the your yield, max, to maximize quality.

[Yeager] Okay. You have to. What did you say you had to focus? so a human eye, is the human eye more effective at, instead of a general cut? Everything this technology is able to, to use a technical eye, it's trained and can see that spear needs to be picked. That spear needs to be peck. And the action goes through. Is that how this works?

[Lu] you. Yep. Yes. the vision system. The camera can see this. Asparagus can also pay down a train. A model can tell a hollow spear how long it is. The selective harvesting is based on length of the skin. We were, say that even the spear longer then were considered as ready for harvesting. If It was a shorter than that, so maybe a little pick next time. So the vision system can do this. The decision making. Tell the spear where to locate it and how if they're ready for harvesting or not.

[Yeager] I'm just fascinated by that. So, let's go back to the cost discussion. That sounds expensive. Is it?

[Lu] Yeah. As a current, in a commercial prototype, these still slow. And also they are expensive. They end, their company based in the UK and the E in Germany, they already have their product available on the market, but the so far, no growers has really adopted their, their product. Yet. There are some issue. There is asparagus. You need us to, to pick them in a very timely manner. You need the legs one day or two days. That's aroma to us. You cannot sell to the girl, so you cannot sell her market. So you need a pick. You need to be very efficient. 

[Yeager] Does this machine operate with a. Is this, is this Wi-Fi enabled that there's somebody sitting at the edge of the field watching this thing go back and forth, back and forth through the rows? Or are they sitting on it like it would be a tractor? to kind of keep this thing going?

[Lu] It's the, the, the linens, the process of the harvesting this and fully autonomous. So no human operator. No operator. Yeah, that's, you can program the robot. So that is a part of the work. so, as I said, there's no then there's what I see. There's one. And then UK based, that's the company they have that is just a single road covered robots, single row on each hand. This, this robot only pick the one spear and it's moved to the next one, so it'll work. But it says what could to slow to be useful.

[Yeager] Now, you sounded like my dad. I was too slow to be useful in the field.

[Lu] Then the grass is it cannot wait. So they want to harvest their crop of this moment. We see it today they need to fully harvest all ready crop.

[Yeager] But the thing with asparagus though I mean any crop I know I'm we're very specific asparagus here, but this really applies to any crop. There's only it's only mature for what, three weeks? Six weeks at most that you need.

[Lu] Yeah. This is then the picking the season can last of all seven out of the nine weeks. Yeah. The picking need to be continued almost every day. So it depends on the weather, all the temperature conditions, depending on how fast it grows. So they're not a very common they were to the availability. all the, even every day or even twice per day.

[Yeager] So this picking frequency is pretty high? In part, is 7 to 9 weeks. So then that doesn't add up to 52. I know you don't use a combine. you know, use a big John Deere or case combine all your through. What is this? What can this machine do in the off 45 weeks a year and is can it be reprogramed for other crops.

[Lu] Yeah. That's where add a value to the machine. But to the so far the robot and to like the robot work very well all the the those current development and they are right task a specific task as simple as NASA restricts their value. But so that's what, research already started. How to like making a robot a multi functioning, kind of value to a robot. But for asparagus so far we haven't really reached that level yet to develop. Very versatile from far farming less, robot. Yeah, but that's good enough. So you.

[Yeager] Can you… Go again. I'm sorry I interrupted.

[Lu] Yeah, yeah. I think you make a very good point that that's the potential to the line of the robot to more than just harvesting. maybe the harvesting. Just the one that can the module that's can be attach or detach. So for all tasks, you can attach the other module, the basic frame can still there.

[Yeager] Okay. Kind of like well like a corn head or a bean head essentially, you know in a very large trying to apply it to my limited knowledge of things here, doctor. So, but because those crops, again, I think California, where we grow. I know you Michigan same way different seasons. You have different needs. And if you've only taken one crop of 300 and solved the technology issue, you still have a whole lot of other things. So, yes, I would imagine your research is endless trying to figure out how to make it most cost efficient for the producer and and would be probably one of those barriers to full implementation. Is that right?

[Lu] Yes. Yes.

[Yeager] In addition to the part of it's got to work to your right.

[Lu] Yes.

[Yeager] Tell me, what are some of the other challenges that you might be facing?

[Lu] You, for vice versa or, development. Right? Yeah. This crop is a very tough crop. I talk with many people. Out of all the folks they has been working on the, this the technology development, I actually, talk to the one, mechanical harvester inventor last year, who made the very early version. That's the mechanical harvester in 1970s. He still had his machine made a separate, iteration. So several generations tried making it work. Still does not work. you know, I mean, it still does not work. I me still does not working as expected. Efficiency. And another issue. So this crop, you know, this, the spears, the crawl in very random pattern. You don't know where a spear will come out. And as they are, sometimes they clustered shoulder one long one. Where nearby. You won't call the shuttle one. You sort of wanted them into this kind of long range. No water damage, that regular shuttle. So this is highly precision, like a harvesting. It's not easy to do the harvesting in a very efficient way. Like I see a lot of the platform traveling at a two miles per hour, that this is just very challenging.

[Yeager] Yeah. And it's asparagus. Takes a while to get to that really nice, width. that it's good. It takes a couple of years. Three years, four years. So if you cut too far or you can do damage in a human can do the same thing. A human could do damage to a plant in the field to not just a machine.

[Lu] Yes. Yeah. So that's the complexity, caused by the crops. So we need to address this before we really can have a, that's practically used for machine. But I think at some point there must be a trade off. So the efficiency and accuracy, there has good trade off. We can even keep this an accuracy estimate of our top priority. Okay. The efficiency, the machine will not be used because there will be too slow to spend a lot of the time and try to find the best cutting angle and to to finish the cutting. That would be too slow.

[Yeager] Right. I hadn't even thought about the cutting angle. Yes, but even a large 250, half $1 million John Deere tractor, combined going through the field will not be perfect all the time. And it will leave, parts of the corn cob in the, in the, in the hopper. So even those machines aren't 100% perfect. They're darn close. But I'm guessing that is that the bar and standard that you're trying to get to is, is what the modern production agriculture machines have done. That's where you're hoping that we can get to, is it.

[Lu] There's some research on the cost benefit analysis. All I don't want a level. this machine can start to generate a profit. So the. If we can harvest the legacy, the is over 75% of the eligible spears. That is, can start to generate the benefits. You don't need to harvest all the crops.

[Yeager] You know, the large combine. Still, they leave parts of the cob in, in the corn hopper. I've seen it for myself. so they're not perfect. But, I mean, your goal, I think you just kind of talked about the intersection of efficiency has to be met at some point. So how close are we? How far are we years? Are we decades away?

[Lu] We are close to the, very commercially variable product, but I cannot give you exactly timeframe. I do see the increased the activities in domestically and internationally on developing selectively harvesting robots for us partners in our neighbor Canada. last month I met a company that was startup company. These, bring their harvest in the prototype to, that's the, we have, events at in Grand Rapids, for the farmers show. So I saw their system on a talk with the president. They are making pretty good progress. Later on, when we have some collaboration. So, I do see this will be coming soon. Things, maybe in the next two to 3 or 5 years. So we may see something. what can what, we're, well, in the field, actually, in the industry. Really need cannot wait a long asparagus industry in Singapore urging the need for harvesting technology. we I feel obligated to make the solution. available next this year, hopefully next to 3 or 5 years.

[Yeager] Doc, doctor, there are so many specialty crops in Michigan, so I'm guessing there's always the hope of another crop. So asparagus is one. What's another one of them specialty that could benefit. Or there's research that could benefit another crop.

[Lu] You mean, the robots?

[Yeager] Yeah. What, where's the research now? What's the next one that might benefit from your work?

[Lu] Oh, I'm working on multiple. The crop. not far from different projects. asparagus and blueberry I pose. So that is also the cherry. So this is major crops. now have projects working for.

[Yeager] All right? Now, those are huge projects for Michigan. I think Michigan blueberries and cherries all the time. So let's quickly talk about cherries and blueberries. What's the challenge in there? In getting the technology right for automation?

[Lu] So blueberry my projects. That's the money. how to estimate the yield before the harvesting. how to determine when is a good time to harvest them? blueberry. Also, they matured in very a uniformly way. You see, the blueberry, they may peak, early seeding only like a tiny fruit. Tim. Blue. And it's still a lot of, like, grain, but they can scale the very first. The peak then will continue for next the pick until is, say, as 70 or 60% of fruits all blue. They would like the mechanical to run once over to clear the crop. So I want to determine when what would be the best timing for picking. So how to do the determination we need to estimate the harvest overall maturity level. We need a count of the fruit alone, like the venous system, to count the fruit, to determine which fruit is blue, which is number to give overall estimation of how much fruit. Tim Bloom This information will be utilized for decision making.

[Yeager] If it's so much on the eye and how much it looks, I'm guessing then are we only going to be able to do this work in daylight hours? Or is it going to be better in night hours because you can control the light source that would help you identify what needs to be picked or not?

[Lu] Yeah. This is actually, we will have the witness system travel in the orchard to scan the canopy. Give this the information, and has scheduled their harvest accrual to schedule their resources for, so the farm management, also will wanted to estimate the yield based on the detection, the blueberry. We can give the estimate that what how much they were able to produce to see that, this will be very valuable information for growers. We already has something like, show to growers. They want to really to see how much how I could we can tell your estimation that's still ongoing work.

[Yeager] There's a new way to stay connected and know what's happening with market to market. When you subscribe to Market Insider, one email and a lot of information awaits you go to MarkettoMarket.org and subscribe to Market Insider.

[Knutson] Hi, I'm Julie Knudsen and joining me today is our host Paul Yeager.

[Yeager] I'm the guest in my own podcast. I always thought this would happen.

[Knutson] You did.

[Yeager] Yeah. Last week after the last time we did.

[Knutson] This, you thought I would be able to take control of this? Okay, I'm trying to get more comfortable in this role, but it is still very foreign to me. But I'm going to do my best to try to lead this conversation and not let you take over.

[Yeager] Easier said than done, I know.

[Knutson] Maybe taken us a couple of tries to get there. We'll find out. So today we do want to talk about the control room though. Do you know what the control room is?

[Yeager] It's where all the magic happens. But that's not important right now. That is it.

[Knutson] It is important because without that control room, we wouldn't be able to do all of the things that it takes to get the show on the air. So in the control room, there are four positions that we have in there the director, the technical director, the graphics operator, and the assistant director. So those positions in there help get the show going and moving and all of that. There are a couple other positions that aren't in there, but we'll dive into those later.

[Yeager] Okay, so those are all the roles that you don't see behind the camera. But if they are not there, we don't have a show. Okay. director, I think in the previous version of this, I called you the conductor. Is that accurate?

[Knutson] Yep. You did. And I think it is a pretty accurate description of what's happening. So the director and technical director, a lot of the times, the two positions I'm doing in that room and as a director, you are cueing the graphics operator, cueing the camera guys, cueing the audio operator so they know when to play a certain, cut a music or hit a and announce or any of those extra things that don't happen all the time.

[Yeager] Are you pushing any buttons?

[Knutson] When I'm the technical director, I'm pushing buttons. When I'm directing, I'm just pushing yours.

[Yeager] How true.

[Knutson] but the technical director does push. They are the ones. Then the director gives all the information, and the technical director then has to take all that information that the director is giving it and then actually execute it. And it's important then that the director is clear with their communication of how they want it to look, whether they're wanting to dissolve to a camera, whether they're willing to take to it, if there's a special effect that they want to use, where the source is coming from for that lead, that we're rolling all of those things very important so that the the team knows, like, okay, this is coming from this source, I have to be ready to in our case, we the TD does have to cue that up even and roll it and.

[Yeager] See what you just said. All I heard was switching. And those words still don't mean anything differently to me, but they do to the technical director. So again, we all speak different languages and that's yours.

[Knutson] Exactly. I have to figure out how to make sure that the TD and the director are on the same page, whatever role it is. So it does help being, having sat at that technical director bench and hearing that knowing like, okay, this is the important information I need to know to be able to accurately execute this.

[Yeager] Okay? This is a total wrench into what we've planned to do, but can someone on television make a director's life easier or harder by not being cooperative?

[Knutson] I think you know the answer to that. They can definitely make it more difficult.

[Yeager] And I can make it very difficult at times.

[Knutson] Well, only when not difficult, but just not prepared. Right. So you might think we've talked ahead of time about a plan that we're going to do. We think we are on the same page. And all of a sudden you throw an audible and say, we're actually going to go to this graphic and said, it's like, oh, so I'm making sure the graphics person is aware that we're not going to that graphic. We're going to a different one. The TD knows that that source isn't going to be playing. It's going to be coming from here. So just quickly communicating very effectively so that where we don't have the wrong thing up on the screen or give any sort of anything wrong that way.

[Yeager] In the first ten minutes of the program, it is very tightly scripted. Everybody knows what's going on. The curveballs usually come in the analysis portion. When we say we usually go wheat, corn, soybeans, live cattle feeders, hogs. Yeah. But there'll be times say it's a show where we have Jeff French and Ross Baldwin, where we're talking livestock and grains all at the same time. And it is just a and you'll say, Paul, where are we going to going on? Be like, I don't know. I just have to know where the discussion takes. I have to listen to where the dance is. Yep. Okay. I think that's a very good description of what a director and technical director. I think that was a really good job.

[Knutson] Oh well, thank you because it's kind of what I do. But you know, what I don't do is close a show.

[Yeager] Okay. Good luck.

[Knutson] Here you go. I'm going to give it a shot. If you have anything you would like Paul and I to talk about, go ahead and send us a message at Market to Market to Iowa. No. At Iowa. PBS.org.

[Yeager] And it's in graphic form. It bails you out there.

[Knutson] Right. But then it may look even stupid or not even knowing what it's saying right here. Market to Market at Iowa. PBS.org. Make sure you catch Paul's new podcast that comes out every Tuesday. And who knows when you'll see us again doing what we do.

[Yeager] Watch every episode so you never miss one of these fun things. Great job Julie.

[Knutson] Thanks, Paul.

[Yeager] All right. See you next time.

[Knutson] Bye.