TerraSentia robot automates the labor-intense process of collecting data to improve crop breeding
Because plant breeding and seed product development is a long, expensive process, researchers want to find any problem a variety might have as soon as possible to get it out of the pipeline.
In the past, measuring traits like plant height meant relying on humans to assess plants. Armed with a ruler and an electronic device to record data, large teams typically made up of high-school-age students would walk through rows of corn as they yelled out numbers.
“It was really hard, because sometimes you’re dealing with corn that is 14-feet tall,” says Neil Hausmann, Corteva Agriscience Field Sensing Lead and Distinguished Research Fellow. “The measurements were not really that good, but it was difficult to collect it any other way.”
To better assess traits, Corteva began employing drones about five years ago. Today, the company’s 500 drones not only assess plant height, which impacts lodging and yield, but other important traits once measured by hand.
“However, there is still a lot you need to know about the crop that you can’t necessarily see from above,” Hausmann says. “As we built our drone fleet globally, we always knew we needed in-canopy measurements, often through imaging, but through other sensors as well.”
Weighing in at 30 pounds and just over a foot wide, TerraSentia is designed to automate in-field plant trait collection, especially under the canopy. Using computer vision and machine learning, the autonomous robot is currently being taught to measure early vigor, corn ear height, soybean pods, plant biomass, and to detect and identify diseases and abiotic stresses.
Developed by researchers at the University of Illinois, TerraSentia is available through EarthSense Inc. To date, 80 robots have been deployed. It plans to produce 100 more this year.
“The quality of our decisions, and, therefore, the quality of our products is driven by both the volume and the quality of the data we collect,” Hausmann says. “Because we don’t have all of the tools to achieve that, we need to partner with a start-up like EarthSense to develop hardware as well as analytics to get the best possible data for our product development.
“Corteva is really a data-driven machine,” he continues. “Plugging into tools like TerraSentia improves the outcome for the grower not only to achieve higher yields, but also to help create more stable products for his operation.”