On a sunny morning in late summer, Precision Agriculture Specialist Hemendra Kumar strolled through an eastern Maryland farm field to test the skills of a new kind of farm worker. This one wasn’t outfitted with muck boots and Carhartts, but with propellers and multiple cameras.
Kumar's farm assistant is a drone.
The drone is connected through artificial intelligence to decision making programs that help farmers assess field conditions and make real-time decisions to improve crop health and yields. “I want every farmer to have Agricultural Intelligence, what I like to call AI, right in their hands through drones, smartphones, and other precision ag technologies,” Kumar said. “To me, Agricultural Intelligence represents the true purpose of artificial intelligence: empowering agriculture with smart, data-driven insights that make farming more efficient, sustainable, and resilient.”
The Future in View
The drones Kumar uses are gaining ground in agriculture, but the algorithms connecting their data to the decisions farmers need aren’t fully developed. For now, farmers use drones like an extra set of eyes, still reviewing images or analyzing data themselves. At UME’s Precision Agriculture Lab, Kumar leads a growing team of graduate students, postdocs, and undergraduates developing high-tech systems that connect the data collection to solutions in action.
His lab is developing AI models that predict soil moisture levels and irrigation needs for corn and soybeans from drone images in real time. These models not only tell farmers how much water crops need but also pinpoint where and when to apply it, eliminating wasteful over-irrigation.
Another focus is evaluating crop emergence after planting. Imagine surveying a 40-acre soybean field two weeks after seeding. Inspecting each row for tiny green plants is time-intensive. With the technology Kumar is building, a farmer can fly a drone and know within hours exactly where crops failed to emerge, and what conditions may have contributed to the problem.
The team has also used surveillance drone data to generate maps that they integrate with another type of drone-one that carries chemicals or seeds and equipment for spraying them. “In the future I envision, farmers will seamlessly manage their fields using Agricultural Intelligence, right from their smartphones or laptops to command drones that survey fields for all kinds of conditions, from water stress to disease and pest infestations,” Kumar explains, “They will then automatically generate irrigation schedules or chemical applications plans with real-time maps that direct irrigation equipment or spray drones to apply the proper treatments where needed."
“The beauty of this AI-driven approach is precision,” Kumar said. “You apply only where needed and you avoid over-application.”
Five Eyes in The Sky
One challenge of Kumar’s work has been collecting the data needed to teach the algorithms what they are looking for. To determine irrigation needs, for example, he has installed soil moisture sensors in farm fields throughout Maryland and used AI to correlate that data with the drone imagery he collects, along with other important information about water availability.
It takes thousands of images to teach an AI model. And these images see beyond what we can see with our eyes. The multispectral drones Kumar uses are equipped with five sensors capturing different wavelengths of light: red, green, blue, “near-infrared” and “red edge,” a wavelength that reveals plant stress earlier than others. Kumar’s algorithms use these images to detect and map plant height, leaf area, and water stress across fields with high accuracy.
They’re now working to recognize plant diseases and pest infestations. The team’s first goal is to develop a phone app or web tool farmers can use without a drone to identify problems instantly. “Right now, if you see a problem, you send a sample to a lab and get results in a week,” Kumar said. “We are developing algorithms that detect disease in real time with a phone scan-it will tell you right away what it is and how to treat it.” Eventually, these instant ID tools can be linked with drone technology, allowing quick, reliable, full-field disease surveys.
Driven To Solutions
Kumar’s enthusiasm stems from long days spent farming and collecting data in hot, sunny fields. He grew up in a small farming community, learning every task firsthand. Later, during his PhD at Auburn University in Agricultural Engineering, he again walked hundreds of acres with GPS to install sensors and retrieve or collect data. He spent days and weeks at a time in the field, away from home in Auburn.
“I thought, there has to be a better way to get this data,” Kumar said. He looked to drones and technology, and since then, he has been working to make life better for others. In a world facing climate unpredictability and finite resources, the stakes for farmers are high. Society needs better ways to farm more efficiently and intelligently to feed a growing population. AI is increasingly providing answers, and for Kumar and his team, the sky isn’t a limit-it’s a vantage point.
At the end of the day, when the drones have landed and the data has spoken, what remains is the promise of better yields, less waste, fewer surprises—and farmers making decisions not by gut or guess, but by the eyes in the sky and the intelligence in the code.
by Kimbra Cutlip
Precision Agriculture Specialist Hemendra Kumar is Training Robot Farmers to Protect Crops and Increase Food Production