AI and satellites transform rice farming

Dec 16, 2024
By Farms.com

New technology helps combat food insecurity worldwide

Researchers from North Carolina State University have developed a pioneering method combining satellite imagery with machine learning to improve rice crop monitoring. The innovation addresses food security concerns and ensures efficient resource use globally.

Focused on Bangladesh, the study tackles traditional challenges in rice yield estimation, such as time-consuming fieldwork and inaccuracies. “It is a time-consuming and labor-intensive process,” explained lead researcher Varun Tiwari. “They do not have this information in time to make decisions on exports, imports, or crop pricing.”

Using time-series satellite data, the researchers tracked vegetation health, water content, and soil conditions. The trained machine learning model delivered yield predictions with an accuracy rate of 90-92%, providing timely data for decision-makers to improve storage, transportation, and crop management

Bangladesh, where rice contributes significantly to the economy and diet, was an ideal starting point. Tiwari emphasized, “If we can get similar data sets from other regions, we can apply this same framework there.”

The research involves collaboration between NC State, USDA, and international organizations. Funded by the Gates Foundation and USAID, it marks a critical step in addressing global food security. Findings were published in PLOS ONE.

Video: The 15-Year Bet Behind Every New Variety!

Canada is trying to decide how much agricultural research capacity it can afford to lose. Brian Rossnagel believes the better question is whether the country can afford to rebuild it.

The longtime barley and oat breeder makes the case with a simple fact about his profession: the consequences of today’s decisions may not become visible for 10 or 15 years.

“Pick the right parents. That’s the biggest thing,” Rossnagel says. “If you pick the wrong parents, you’re not going to get anywhere—and you don’t know that until 10 years, 15 years later.”

That warning carries particular weight as Agriculture and Agri-Food Canada moves to reduce spending and streamline parts of its science operations. The department’s 2026–27 plan anticipates the loss of approximately 665 positions by 2028–29 and says some research will be reduced where capacity exists in academia or industry. AAFC says the changes will make its science operations more cost-effective over the long term.

For Canada’s seed industry, Rossnagel’s career illustrates what is at stake.

This fall, the retired University of Saskatchewan breeder will be inducted into the Canadian Agricultural Hall of Fame. During his 35-year career at the Crop Development Centre, he helped develop more than 100 barley and oat varieties, including CDC Austenson—one of Western Canada’s most widely grown feed barleys. His induction recognizes not only those varieties, but the collaboration and research system that made them possible.

Rossnagel is quick to emphasize that none of it was the work of one person.

“The first thing I thought about was all the other people who contributed to whatever success I and my program had over the years,” he says. “We know that it’s not an individual who does this. It’s a group—a team.”

That team extends well beyond the breeder whose name appears beside a variety. It includes technicians, pathologists, quality specialists, statisticians, regional testing sites, seed growers and industry partners. It also includes the breeders who came before and those who will carry the germplasm forward.

CDC Fraser barley, for example, moved through three breeding careers. Its parents came from Brian Harvey’s program. Rossnagel advanced the material after Harvey retired, and Aaron Beattie later guided it through registration and release.

That kind of handoff is normal in plant breeding. The person who makes the original cross may never see the resulting variety reach farmers.

It also explains why lost research capacity cannot simply be switched back on when budgets improve.

“If you shut it off, it’s very, very difficult—and particularly costly—to start it up again,” Rossnagel says. “If you have to start from scratch, it’s going to be at least 10 years before anybody notices whether you’re getting anything done or not.”

The concern is not simply how many experimental lines Canada can process. Modern equipment, statistical tools and genetic technologies allow today’s breeding programs to evaluate tens of thousands of lines—far more than Rossnagel could handle when he entered the field in the early 1970s.

But efficiency and automation do not generate every idea.

“If you pare back down, and instead of having six or seven individual scientists concentrating on wheat breeding, you go down and say three people could handle all this, well, that’s half the ideas gone,” he says. “Particularly if you happen to lose the three people who had the really neat and innovative ideas, boy, that’s a problem.”

It is a timely distinction for Canadian agriculture. Consolidating programs may preserve the volume of material moving through a system, at least initially. It may not preserve the diversity of thinking, regional knowledge or willingness to pursue unconventional crosses.

That regional knowledge matters because Canadian agriculture is not one uniform production environment. A variety suited to southern Alberta may face different disease, moisture and maturity pressures than one grown in Manitoba, Ontario or Atlantic Canada.

“Agriculture is applied biology,” Rossnagel says. “Biology, all around the Earth, moves from the poles to the equator. It does not move from Newfoundland to B.C. like politics do.”