AI in agriculture: Augmenting farms with object recognition
Dan Vertachnik from Twenty20 Solutions explains how his team is using AI and IoT to improve California’s cannabis harvests
Dan Vertachnik from Twenty20 Solutions explains how his team is using AI and IoT to improve cannabis harvests in California
When it comes to stereotypes, the notion that farmers have to be armed is as old as time itself. The clichéd 'get off my land' wouldn’t work without a shotgun in hand.
And in some instances, that stereotype remains true – farmers must protect their vast, remote fields, their livelihoods, often alone. Is there a way technology can help?
It can, with remote monitoring systems and image recognition to track vagrants and trespassers to determine potential threats.
AI Business spoke with Dan Vertachnik, CEO of Twenty20 Solutions, one of the companies working in this space.
Vertachnik explained how his team, traditionally a security technology provider, entered the agriculture market, and why he feels AI could revolutionize the farming world – in terms of both safety and efficiency.
You say tomato, I say potato
Vertachnik and the team at Twenty20 entered the agriculture market around 2015.
“We had a farm come to us that was having problems with tomatoes on the conveyor belt – determining the right size, ripeness, color – we put in security cameras to be able to monitor what was going on,” he explained.
Since the original deployment, the contracts have been coming in non-stop. Today, a sizable portion of the Twenty20’s teams agriculture work is in California – a state that legalized cannabis in 2016 via Proposition 64.
Legalization opened new doors for the company. Shortly after, a cannabis farm sought to secure their manufacturing facility, and Vertachnik answered the call, installing door and gate access, as well as smart security.
“We then looked at fields – and because farms are remote, they don't have a lot of electrical power capabilities.
“Our strength was we started our business by servicing harsh off-grid environments, remote environments where you didn't have great cell coverage or the capabilities to put a power line in. We would use our solutions which are powered by solar, batteries, generators.”
“Since entering [the agriculture market], it's grown dramatically for us over the last seven years,” he added.
Object recognition and automated irrigation
Agriculture isn’t the only new market for Twenty20: it has deployments monitoring everything from cell towers and data centers to residential properties and military equipment.
What has Vertachnik learned from working with farmers? Adhering to the unique requirements of operators in this space, he replied, adding, “I think we learn something cross-vertically all the time.”
"Among our work, we manage the security inside hothouses. When you deploy inside a hothouse, it's warm, it's humid, and the equipment keeps dying. We had to come up with revised equipment which we built especially for the industry and now we can do hothouses. Now we're using that across other, different segments where they have heat issues.
“Cannabis growing laws in California state that not only do you have to record and store things in the cloud, but you have to store all motion from all cameras 24/seven for 90 days. it's a unique equation and we're able to do that.”
Vertachnik’s team also works in automation – using sensors to measure valves and tanks in places like factories and power plants. The company took some of these concepts and applied them to farming, using sensors to determine soil moisture levels to automatically turn on sprinklers or irrigation systems if crops become too dry.
“Last year, we released technology for object recognition and classification that's really strong, because in agriculture, they're in the middle of a field with animals roaming and tumbleweeds flowing through. You want to be able to say 'hey, that doesn't have two legs so it is not a threat.'
“We take that security and technology into the field.”
Analytics and agriculture
Automation and AI deployments in farming have come a long way since Twenty20 first stepped into the market.
“They're looking to use the data to plan a year ahead and want to build historical data. It's becoming very, very sophisticated,” Vertachnik said.
“Sensors – that's always the first question 'can you take the readings? Can you provide us information from the data we can use?' It's a very exciting industry right now.”
Today’s farmers have to contend with a myriad of issues, including the effects of climate change like flooding and droughts, worker shortages, and customer demand fluctuating amid the pandemic.
Vertachnik suggested that the industry is looking to technology for help: Farmers are all saying 'what can you do to help us on our processing and productivity – can you help us reduce our costs and improve our margins. We're seeing a lot of questions like that coming in.”
One area his team is obviously focused on is site security.
“They're trying to control access to facilities, onto the properties themselves – we're doing a lot in that area right now, automated access control to say who is at the gate, and whether they should come in.
"We have technologies that tell farmers whether a truck belongs to you – checking logos, license plates – using AI, we look at the type of vehicle and the length of time it spent on the land.”
Automation and asset management
The demand for intelligent monitoring is increasing. Vertachnik revealed that his company doubled in size this year and “expects to do that again next year through acquisitions, as well as organic growth.”
“This is one of our two fastest-growing spaces [for us]. Utilities and agriculture are just going crazy right now. In the cannabis region, we've become a very strong player – every quarter we're signing new cannabis customers.
“Now we've entered the apple market – I see us growing. We're poking around and getting interest from folks growing avocados, I think they happen to be green gold right now.”
Twenty20 is now looking at expanding their product portfolio.
“In the next several months, we're looking at bringing asset management technology into the mix so farmers can monitor their products,” Vertachnik said.
"We think that optimization and automation is the future of the agriculture industry. Automation is nothing more than workflow management and optimization and then managing the data to be able to access the controls. Every day we're looking at what can do in this space to help improve productivity and the process – and to manage those costs."
AI and IoT
Vertachnik seemed positive that AI and emerging tech will go a long way to helping farmers in the next five years.
“When you combine AI with an IoT platform you get a very powerful tool,” he said.
“The farmer does not want to sit all day and watch video to determine issues. We can set it up so there are 100 cameras, but just one screen showing a unique event that tells them when something is wrong.
“Farmers are looking for increased ability to manage without increased cost. Automation and field management are the things that are going to continue to grow.”
This interview is part of AI Business’s focus series exploring the role of AI in farming and agriculture.
Over the next few weeks, we’ll be interviewing innovative startups and hearing from some of the market’s biggest players to find out how AI and computer vision are changing agriculture for the better.
Check out our market overview, as we ask: can AI help save the farming industry?
About the Author
You May Also Like