Technology of the Year: A New Vision for the Future of Greenhouse Production

FarmVisionAI camera by IllumitexIn July, Greenhouse Grower® magazine announced Illumitex’s FarmVisionAI visual imaging artificial intelligence (AI) platform as the 2020 Technology of the Year Award winner. Part of the Greenhouse GrowerSM Medal of Excellence program, the award recognizes new innovation in structures, automation, or equipment that makes controlled-environment crop production more efficient while improving plant quality (check out this video highlighting all of the finalists for the Technology of the Year award in 2020).

How it Works

The FarmVisionAI system, launched at the beginning of 2020, includes five elements:

Advertisement
  • High-resolution, wireless cameras that can cover 16 to 100 square feet of plant canopy, with configurable lenses
  • A ubiquity network that routes data gathered by the cameras
  • An ITX Edge server that stores images and is easy to install, with a high security gateway to manage access
  • A secure cloud storage system
  • A software user interface

FarmVisionAI works in both a web-based app and a mobile app. The web app has dashboards based on visually derived, real-time output data, with alerts on nutritional deficiencies and critical success indicators. The mobile app allows growers additional features when connected inside the greenhouses, with location-based problem identification and solving. Growers can add new images and can make quick scouting notes while inside the facility.

The system collects images and incorporates them into an algorithm framework that applies deep learning to quickly show the grower hot spots for poor plant quality. FarmVisionAI also has a tier of service that includes remote horticulturists examining the imagery to further identify issues that range from identifying botrytis, powdery mildew, and mites, to problems as simple as knocked over plants, poor post trim clean-up, or a clogged emitter.

Benefits and Education

According to Jeff Bisberg, CEO of FarmVision AI, by Illumitex, the goal of developing the platform was to use the latest technologies to provide new real-time digital tools to improve the overall productivity of controlled environment agriculture.

Top Articles
The Benefits of Automating Your Cannabis Greenhouse

“Farming at scale is very difficult, and we started with a simple proposition: how could we ensure that every single plant in the greenhouse was thriving?” Bisberg says. “If we made sure that every plant was healthy, then the farmer’s production quality and yield would increase, and they would become more profitable.”

While most agriculture technology is focused on inputs and processes, Bisberg says nothing was telling the grower what was happening with the most important part of the greenhouse — the plants.

“We aim to fill that missing link by giving the farmer real-time information and advice on each and every plant,” Bisberg says. “The grower can use that data to drive all the details required to produce the best plants possible, and business improvement will follow.”

As FarmVisionAI launches, Bisberg says the initial focus is on high-value crops, so the grower can get a quick return on investment.

“High-value determinant plants are best for our initial camera-mounting strategies,” Bisberg says. “We are targeting cannabis, but we are also seeing the potential for use in some specialty plants.”

To address the steep learning curve involved in AI technology, FarmVisionAI is developing a certificate program on their websitewhere a grower can learn and get certified as a Computer Vision Horticulturist.

“Education is the best way to make growers and farmers comfortable with new technology, whether it’s hardware or software,” Bisberg says. “The training consists of educational material on how computer vision is applied in horticulture, and has a series of quizzes and tests that must be passed to get certified.”

Why Illumitex Went from Lighting to AI

Growers in horticulture may be more familiar with Illumitex as a lighting supplier. So why did the company make such a dramatic shift?

“When I was out on sales calls, I quickly learned that lighting, although the most important input for photosynthesis, was only a small part of farming and growing plants,” Bisberg says. “We have great horticulturists on our team who are highly educated, and they would also advise on everything from nutrition to environmental monitoring to irrigation and more during those sales calls.”

Bisberg says it became obvious that the advising part of the visit was creating differentiating value, building trust, and opening up new opportunities.

“It was this need to scale horticulture advice, combined with my previous experience with VisionAI, that led to the start of FarmVisionAI to solve the problem of deploying horticulture science at scale.”

At the time, Illumitex as a lighting company was growing at close to 50% year over year in the horticulture space, Bisberg says.

“Unfortunately, the number of lighting companies seemed to be growing much faster, and that included many low-cost providers from China,” Bisberg says. “To be frank, it’s just too easy to build a pretty good horticulture LED light fixture and pitch it like a great fixture. Many customers don’t have the time or interest to learn the details, and that led to intense market competition and dropping prices, which is good for the grower but bad for the manufacturer.”

Concurrent to those discoveries, Bisberg says the FarmVisionAI platform was gaining traction with customers.

“Late last year, we made the call to radically change directions to a pure play VisionAI ag tech company,” Bisberg says. “When COVID-19 hit, it only solidified our resolve, as our VisionAI system really addressed the issues of remote work and visualization on the farm head-on.”

Next Steps

The initial rollout of FarmVisionAI has been met with positive feedback. So, what’s next? Here are four design elements the company is planning:

  • A mobile app that enables real-time, location-based interaction in the field. This helps the grower solve problems that are more complicated than AI can detect from above, Bisberg says.
  • Increased number of detections of both good and bad things happening in the greenhouse. “We try to push our detections to the very first moment a defect is detectable and make that detection robust enough so that we can make meaningful changes early in the grow,” Bisberg says.
  • An open interface so other farm software providers can interface to the data to make their software more powerful. “We realize there are a number of providers solving important problems on the farm, and we want to partner with them for more comprehensive solutions,” Bisberg says. “We sort of think of our FarmVisionAI as the Air Force, but you also need the Army, Navy, Marines, and National Guard to fight the battle and create a profitable farming business.”
  • More kinds of cameras that can work with vine crops and thermal imaging.

1