Dr. Chris Reberg-Horton is a self-described “radically impatient person.” Turns out impatience isn’t always a bad thing — sometimes, it’s exactly the trait needed to ensure that food makes its way from the farm to your table.
A land-grant university is just the kind of work environment where his impatience thrives.
“It fits my personality. I like working at the edge of advances and getting them directly into farmers’ hands, faster,” says Reberg-Horton. “People are often mesmerized by the idea behind land-grant and extension universities, and it’s part of what makes NC State stand out when you think about institutions worldwide.”
Dr. Reberg-Horton might initially have graduated from the blue school down the road, but he found a home at NC State when he pursued his Ph.D. After a stint working in Maine, Reberg-Horton made his way back to North Carolina to be closer to family.
Now, the Blue Shield of North Carolina Foundation/W.K. Kellogg Distinguished Professor in Sustainable Community-Based Food Systems works diligently to put precision farming back on the table. Reberg-Horton has always focused on sustainability and agriculture, yet in recent years his interests have turned toward how AI can help farmers make decisions.
“I realized that we’ve made a ton of breakthroughs in terms of managing farms, but the answers were complex,” said Reberg-Horton. “Those complexities make terrible fact sheets, and I want to help people put the information into practice.”
This mission pointed him in the direction of software development for farmers, specifically developing computer vision and artificial intelligence to help oversee the health of crops. Farming used to require the precise management of smaller plots of land, but farm consolidation and a labor shortage have made keeping pests, fertility issues and disease under control increasingly difficult.
Reberg-Horton’s team of faculty and graduate students within the N.C. Plant Sciences Initiative (PSI) have developed a camera system to bring precision farming back to the forefront. Mounted on spray booms that are already moving through farmers’ fields, these cameras can identify disease, fertility, pests, weeds and other issues.
“We don’t have the knowledge per acre we need to effectively manage agricultural challenges,” says Reberg-Horton.
Artificial intelligence is helping move that needle back in the right direction. On large plots of lands, farmers struggle to quickly identify problems across their entire acreage, but cameras can be the extra eyes they need to stay on top of agricultural challenges.
“We’ve spent decades treating huge swaths of fields the same, but we have the technology to manage fields in more precise and intentional ways,” says Reberg-Horton. “These cameras can help farmers have a dynamic understanding of where things are going poorly or well.”
One example of this? Cover crops.
Cover crops are plants used to blanket gardens and farms in order to re-nourish the soil. They have a tendency to be patchy, particularly the mixed species plantings that are most commonly used, explained Reberg-Horton. With computer vision, we can map each species in the field and reduce our nitrogen applications where legumes have performed the best.
Reberg-Horton is also researching weed mapping through computer vision technology — as he explains it, North Carolina has a particularly potent problem when it comes to herbicide-resistant weeds. Palmer amaranth and ryegrass are some of the biggest threats to soybean and wheat crops in the state.
“What we are doing is teaching the artificial intelligence program to recognize these weeds within the field, which is a hard computer-vision task. Sorting out one grass from another is really tough, but it’s doable,” says Reberg-Horton. “This information can help growers plan for their next growing season and think about rotating that field or where more herbicides are needed.”
The agricultural potential for this kind of computer vision program is extraordinary, but training it is a labor of love. The program has to know what every problem looks like in order to identify it, so Reberg-Horton’s team is carefully engineering those potential issues on real plants as well as digitally manufacturing images themselves to train the artificial intelligence.
“There’s a robot outside of the Plant Sciences Building and right now we are introducing dicamba to soybean plants in very precise amounts to engineer the exact symptoms we’re looking for in the pot.”
That robot takes pictures of all the plants several times a day, and these photos train their computer vision program to identify dicamba damage on crops that are not resistant to the herbicide, a particular branch of Reberg-Horton’s research taking place in partnership with the US Soybean Board.
Reberg-Horton looks forward to new developments within the artificial intelligence sphere in the coming years, especially as a sustainability- and conservation-focused researcher.
In addition to being the Blue Shield of North Carolina Foundation/W.K. Kellogg Distinguished Professor in Sustainable Community-Based Food Systems, Reberg-Horton co-leads the Precision Sustainable Agriculture organization. He is also the platform director for resilient agricultural systems within the N.C. PSI.
Working with both the N.C. PSI and the Center for Environmental Farming Systems (CEFS) puts a new lens on Reberg-Horton’s work, bringing together interdisciplinary teams that solve complicated problems.
“CEFS specialty is bringing in people from non-science disciplines and helping people like me see the bigger picture. As an applied agronomist, I’m locked in on finding the technological solution but CEFS takes a more comprehensive approach,” he says.
The funds for his distinguished professorship come in through CEFS, and Reberg-Horton says the CEFS team takes a fresh approach to the use of these funds.
“I know professors in other departments use the funds from these positions for their own research, but we take a communal approach,” says Reberg-Horton.
The team at CEFS views the added funds as belonging to the center as a whole and uses a transparent decision process to decide where to allocate the money. Usually that means funds go toward work that will keep the team inspired, as well as projects that continue to provide work for graduate students.
“I find this group of young people inspirational and they really make me think about things more broadly than I generally do,” Reberg-Horton says of the roughly 15 students typically working within CEFS, and they try to give them a real work-life experience while they are still on campus.
Through his work as a platform director in the N.C. Plant Sciences Initiative, he has been able to be a stronger collaborator across campus with other faculty members like Lirong Xiang, Edgar Lobaton and Amanda Hulse-Kemp.
“A lot of what I do depends on engineering, and if you look at our lab, the majority of our team works in electrical or computer engineering and computer science. Being able to collaborate within this building has been fabulous for our work,” Reberg-Horton says.
The work that goes into training this kind of technological system is extensive and often cost-prohibitive for the private sector, so Reberg-Horton is excited to open-source the hundreds of thousands of images his team has collected to reduce these barriers.
“We can open-annotate data sets that will help both private and public developers, which will welcome more innovators to the table to work on problems facing agriculture.”
Right now, farmers across the country are testing the cameras on their fields. Hopefully one day soon, the program will be more widespread, helping farmers grow better, bigger and healthier crops — because as Reberg-Horton will tell you, “that’s just what we do at a land-grant university”.
This post was originally published in Giving News.
- Categories:
