Farmers Use Remote Sensing, Big Data, AI to Simulate Real-World Crop Production Scenarios

By Blair Fannin

Crop farmers in South Texas are witnessing the future of agriculture unfold with the advent of digital-twin technology. Spearheaded by Texas A&M AgriLife Research, this cutting-edge approach combines remote sensing, big data and artificial intelligence to simulate and predict real-world crop production scenarios.

Juan Landivar, Ph.D., director of the Texas A&M AgriLife Research and Extension Center at Corpus Christi, leads a multidisciplinary team of experts, including agronomists, computer engineers, electrical engineers and civil engineers.

He recently shared their findings at the Texas Plant Protection Association Conference, emphasizing this technology's transformative potential. Their methods and results have been published in the peer-reviewed journal Computers and Electronics in Agriculture.

The birth of an idea

The concept of digital-twin technology in agriculture emerged from a conversation six years ago between Landivar and his then-colleague Jinha Jung, Ph.D., now an associate professor at Purdue University.

"We were returning from a meeting when the idea clicked," Landivar recalled. "I couldn't sleep that night. By 3 a.m., I was texting Jinha, realizing the vast opportunities this technology could unlock for agriculture."

This sparked a series of trials on a 200-acre farm in South Texas, cultivating cotton and sorghum, which have showcased the technology's promise. Using drones, the team gathered more than 250,000 data points in a single season, measuring canopy cover, plant height and vegetation indices via normalized difference vegetation index, NDVI.

The challenge then became how to interpret this massive data trove.

Power of AI

"That's where our AI-powered web-based modeling comes in," Landivar said. "It translates complex datasets into actionable insights for farmers, helping with decisions on yield prediction, biomass estimation, crop termination and irrigation scheduling."

One notable success involved advising a farmer to prepare for harvest earlier than expected. In the 2024 cotton crop, AI modeling accurately predicted optimal harvest preparation as early as June 18.

"The farmer said 'no way. I usually defoliate in July,'" Landivar recalled, "but field observations on June 24 confirmed the model's accuracy."

"Somewhere along there, they had several inches of rain and delayed defoliation," he said. "But while waiting for the soil to dry, heavy rains from an approaching hurricane came through and dropped another 4 inches. Harvest wasn't until late July, losing quality and about $70 per acre in potential profit."

Benefits for farmers

Digital-twin technology is ushering in an era of prescriptive agriculture, where decisions are data-driven rather than guesswork. For instance, early yield forecasts—available six to eight weeks before harvest—can aid financial planning and market strategies.

"This precision saves costs and maximizes harvest potential," Landivar said. "It also supports sustainability goals, like estimating biomass for carbon credit markets."