When people picture a working farm, Harmandeep Sharma said most tend to have one image in mind: a farmer in a field of crops, working in soil all day.
While that isn鈥檛 an inaccurate perception, it鈥檚 also not the full picture. In Sharma鈥檚 agricultural lab at North Carolina A&T State University, those rows of crops are embedded with sensors, and drones are flying overhead, monitoring the fields.
And Sharma鈥檚 hands aren鈥檛 full of soil 鈥� instead they are gripping a laptop, inserting data about the crops鈥� health that day.
鈥淭here has been a very rapid advancement of agriculture over the past couple of years,鈥� Sharma said. 鈥淲ith this technological advancement, we have sensors, we have drones. We are in the era of Agriculture 5.0.鈥�
Sharma said the newest venture in this era of data-driven agriculture is artificial intelligence.
Last year, the U.S. Department of Agriculture led by the University of Minnesota. It鈥檚 one of several AI agriculture projects the federal government has .
In North Carolina, the General Assembly between N.C. A&T, North Carolina State University, and SAS, an analytics firm. In March, SAS deployed an AI platform for the two universities鈥� agriculture departments.
John Gottula is SAS鈥� principal advisor for AI and has a background in crop science. He said the AI platform helps researchers make crisper conclusions faster.
鈥淚t breaks down that silo between people who know the subject matter and people who know the data science,鈥� he said. 鈥�(Subject-matter experts) are now able to query that data and crunch those numbers in a meaningful way, where they can quickly recognize what鈥檚 realistic and what鈥檚 not.鈥�
The platform is being used by 14 programs at N.C. A&T, including Sharma鈥檚 Plant Sensor Lab.
Sharma鈥檚 research is on fresh market tomatoes, industrial hemp and hot peppers. She has two lab locations at the A&T farm in Greensboro: an environment-controlled greenhouse and an outdoor field with naturally changing conditions.
Plants at both locations are attached to sensors that continuously collect data. They measure conditions like air temperature, carbon dioxide levels and water uptake. The information is fed into the SAS AI platform, which Sharma said can make crucial predictions.
鈥淚 have the capacity to generate hundreds of models at one time,鈥� Sharma said. 鈥淎nd then ask artificial intelligence to compare those models and provide me with the champion model, the best predictive model for any data set.鈥�
The models suggest optimal water levels, sunlight or other conditions for the plants to thrive. Sharma then uses these hypotheses to help guide experiments in her research.
Before this technology was available, Sharma said she and other researchers had to do 鈥渂lind average applications.鈥� Researchers would decide on one rate of nitrogen or water level to apply on the entire field.
Now, Sharma said she can use technology to map out her field and tailor precise amounts of water, nitrogen or fertilizer her plants need. This practice is called . It鈥檚 a type of sustainable farming that uses data science to reduce the resources used to grow crops.
鈥淲ith this technology, we are using the required amount of inputs,鈥� Sharma said. 鈥淚t's not just environmentally sustainable, because we are not polluting our waters, but it is also helping farmers with the money. They are not overspending on fertilizers that are not helping their crops in any way.鈥�
The point of the collaboration with SAS and NC State is to help local farmers. The universities will share research findings through their cooperative farming extensions. Gottula said that advanced technology like AI usually benefits commercial or large-scale farmers.
鈥淎nd so by working through those institutions (N.C. A&T and NC State), we are able to help bring the benefits of AI, those crisper decisions, that increase the value of farming products,鈥� said Gottula from SAS. 鈥淲e're able to get that in the hands of growers, who may be smaller and less resourced than those with tens of thousands of acres.鈥�
In Sharma鈥檚 lab, her focus is to help small-scale and historically underserved farmers grow high value crops like tomatoes and hot peppers. She plans to condense her findings into easy-to-understand tips that farmers can use to maximize their yields.
鈥淲ith the power of AI prediction, we can tell (farmers) maybe go for this particular variety, this particular amount of inputs,鈥� Sharma said. 鈥淭hat鈥檚 going to save their time, save their money and then it鈥檚 going to help them grow crops sustainably.鈥�
