Modern marvels

As labor availability and costs continue to afflict the ornamental industry, growers dream of automated solutions for production tasks. Last fall in the Greenhouse Management State of the Industry survey, 55% of respondents said they’d like to automate more tasks and 19% said they’ve adopted automation to help combat labor shortages. Much of the latest production technology available to growers winds up being deployed in the cannabis and produce markets.

But there’s hope for the ornamental sector. Greenhouse production in general is moving toward a more data-driven, scientific management approach, creating opportunities for innovation. Public and private organizations continue to research and introduce technology that can — and will — be applicable to all indoor horticulture segments.

Vineland Research and Innovation Centre (Vineland) in Ontario, Canada, specializes in horticultural innovation. In early 2023, Vineland conducted stakeholder interviews with 26 companies and organizations involved in horticultural automation. Interviews included 22 automation companies, three consultants or government representatives and one individual from academia, and these stakeholders represented Canada, the Netherlands, the United States, Japan, Singapore and Israel.

From these interviews, Vineland identified six main challenges facing horticultural automation development.

1. Highly variable and harsh environments are a unique challenge within horticulture:

a. Variability in crop type and variety, production practices, infrastructure and variability over time require more sophisticated equipment.

b. It can be difficult to define an average operation, which leads to the constant need to customize.

2. Balancing the high cost of development with ROI:

a. High variability means large data sets and extensive data collection times are required.

b. Horticulture is a niche industry with each crop and/or growing operation design having unique requirements.

d. Expensive, robust electronics are needed for harsh environments.

3. Developing connections with growers during the development phase:

a. Many growers are not willing to participate in trials due to potential liabilities, high financial cost and time commitment.

b. Word of mouth and personal relationships are key to developing horticultural connections, making it difficult for new entrants in the market to build trust.

c. Balancing the timing of going to trials to demonstrate efficacy versus commercial readiness.

4. Grower buy-in at commercialization:

a. Cash flow is limited, so technology needs to demonstrate an ROI.

b. Technology is not seen as necessary or doesn’t fit existing infrastructure.

c. There is a lack of access to technical support if technology should fail.

5. Lack of collaboration:

a. There is limited horticultural knowledge sharing and a lack of public data sets.

b. Established companies may not be willing to share horticultural connections with new entrants, thereby blocking market access.

c. There is a lack of inter-operability and modular systems.

d. Grower fatigue due to lack of streamlining trials across different companies.

6. Safety and logistics:

a. Biosecurity.

b. Employee security.

c. Compliance with industry and government regulations.

When asked about research needs, the participants indicated they foresee vision systems and artificial intelligence (AI) to be adopted universally in the short term and robotics in the long term to meet the needs of the horticultural farm of the future. Current barriers to universal adoption include cost, technical challenges and lack of collaboration, according to Vineland and the participants.

Part of Vineland’s strategy to encourage and develop more universal adoption of technology includes bringing Brian Lynch on board as the new director of horticultural technology solutions in the fall of 2023. In this role, he will lead Vineland’s “refreshed vision for the Horticultural Technology Solutions program, a hub for testing, validating and optimizing both Canadian and international technology solutions for the horticultural industry.”

“Since his start at Vineland, Brian has shown a strong understanding of horticulture, Vineland’s value proposition and a dedication and passion for solutions in the industry and Vineland as an organization,” says Darby McGrath, vice president of research and development at Vineland.

Lynch joined Vineland in 2018 as senior research scientist and has led a team of engineers and technicians on projects tackling R&D problems in horticultural automation and robotics. He has also been instrumental in project technical developments in machine learning, computer vision, mobile robotics, manipulators and mechatronics.

Greenhouse Management asked Lynch about growers’ needs regarding automation, the impact of costs and more.

GM: Which projects have you focused on since joining Vineland?

BL: I think we’ve been building out a really good sense in our team of how to apply our expertise in that area of horticulture, in the greenhouse. Just because a lot of us are coming from engineering backgrounds don’t understand a lot about the biology [of plants] to start with. There are maybe two universities out there who are graduating people in engineering programs with an agriculture focus. So making those strides to build up a team who can do that is actually quite a big deal.

The projects we’ve worked on, we’ve been able to commercialize two out of three flagship projects we’ve just finished working on for five years. One was around mushroom harvesting using Vision AI robotics. Another one was about AI systems matching grower irrigation decisions in floriculture. And then the third one, which we are working on commercializing now, is around robotic cucumber harvesting for high-wire crops in greenhouses. So a lot of success over those past five years in those technologies, building up that momentum and expertise and honing in on what I found is not necessarily unique to this industry, but is critically important that I think a lot of start-ups are not paying really enough attention to. And that’s the economics around implementing these technologies.

When I first joined Vineland, I went around to growers and asked them if they like technology or if they are scared of technology. And they all kind of said the same thing generally, or at least gave me the same impression which is that they don’t care. If it helps them, they are open to it, they like it. It’s another piece of equipment to them. But at the end of the day, it’s a business. So what they care about most is if they are going to get their return on investment, are they going to get the value of it and does it actually make their life easier.

I bring this up because, in academia, you’re often focused on just solving a problem. You don’t necessarily care about how expensive the technology is or how fast it is or how usable. The big thing we have learned is that it’s critical from the very beginning to understand how to strategize a solution, so you actually end up with a product at the end of the day that is useful and provides that value.

And that’s tricky. So we are right on the verge with those technologies where robotic systems are less expensive, computers are faster, cameras are everywhere. These things are converging, but it’s not quite hands down, for sure you can make a robot do something like pick cucumbers and it’s going to pay itself off in six months.

GM: How far away is that technology for greenhouses from being something with a solid ROI and clear in terms of what it can do?

BL: I think it’s already here. There are some companies that I’ve seen that are doing this the right way. They understand the grower as a customer. Some companies are treating it like labor where they expect growers to pay a variable cost and so there’s a monthly charge for it. It’s not even because they want to do it, I think they need to do it that way because otherwise the payback period is something like 10 years for some of these technologies that are out here being touted at trade shows. In my experience, it’s hard to sell growers an equipment on that monthly cost. ...

But there are companies, and I think we are one of them, that think of technology as equipment, where you can see these things really making sense as an actual product. There are products for things like UV treatments already out there, but not necessarily for harvesting yet. ... The companies developing them just really need to focus in on that price tag and the business model for the grower. It’s not far off. It’s very, very close.

GM: When you talk to a grower about staying up to date on automation and being ready to embrace it when the time comes, what goes into that conversation?

BL: Honestly, it’s the other way around a lot of the time. They ask when it’s going to be ready. They know this technology is out there and being touted. They’ll say, even sarcastically, ‘Where is that robot? Twenty years ago, I saw something in The Netherlands that was supposed to be here in five years.’ So they know what’s coming because they pay attention. They read articles. So they ask more than researchers or start-up industry members telling them to be ready for it. They are the ones who are telling us about the labor crisis they are facing, whether it’s the cost of labor, whether it’s access to staff. In Canada, and the States as well, people don’t want to do these jobs. I think it can depend on the location.

Where we are in Leamington, Ontario, there’s a well-established community of temporary foreign workers who come in from countries like Mexico and Jamaica and so on. But in other spots, those communities aren’t established, so it’s hard to get people to come there to do the work and they struggle to find locals who want to do it.

We don’t have to tout much about growers being ready for it. It’s about encouraging growers to understand that developing these technologies is hard to do in an office, in a warehouse. We have a lab on campus where we’ve got fake cucumber vibes that we got from some place in China. We can’t do much with that. We need real crops to test on. At Vineland,, we have that capability because we have research greenhouse space. We can grow our own. We can test them. And if our robots damage a crop. That’s ok because it’s not our business to sell the fruit.

So getting growers to understand that we need them to be part of the development process is really fundamental. A lot of growers really do get that. But at the end of the day, they are running a business. They can’t afford for a company to come in and break things. And they don’t necessarily have the time to be intimately involved and advising the development process. Even the stuff that’s well established is hard because it’s big numbers. It’s a big check to cut. That’s where the conversation goes — be patient, we are getting there and that when you do see it, it’s not going to work 100%. You’re always going to have staff. But if you can get 80% of the fruit off the plant, that’s probably going to be a great business case for a robot.