Collecting precision crop data with Unmanned Aerial Systems has taken off in recent years. Drones are ideal for gathering data across a wide area in a short time. But to get more detailed plant measurements quickly, researchers are moving underneath the crop canopy and using Unmanned Ground Vehicles to take measurements up close.
Girish Chowdhary, Ph.D., is an assistant professor of agricultural and biological engineering at the University of Illinois at Urbana-Champaign. His lab developed the TerraSentia robot with support from the U.S. Department of Energy’s Advanced Research Project Agency-Energy. He co-founded the robotics startup EarthSense, Inc., with Chinmay Soman, the company’s CEO.
“Aerial systems are really good at collecting a lot of data over a wide scale. Above the canopy, you can very quickly find problem spots with aerial systems,” Chowdhary said. “But when it comes to data under the plant canopy, aerial systems really don’t have access to that because it’s occluded by the plant canopy. That’s where the robot comes in,” he said.
Aerial systems do overall scans of the field. “Then the robots can go in and be the boots on the ground,” Chowdhary said. UGVs like TerraSentia also serve as eyes on the ground for growers.
The 2019 advanced prototype of TerraSentia uses visual cameras, LIDAR (or Light Detection and Ranging, which uses light pulses to measure distances) and other sensors to collect plant data. Chowdhary said two cameras are on the sides of the machine, and another faces the front. A fisheye-lens camera is on top of the robot, and LIDAR sensors are mounted on the front and back. Optional RTK GPS and additional sensors are available for further customization of the robot.
Users control the robots with a tablet application. Software analyzes the data to predict crop performance and yield and extract other trait information.
This year early adopters are continuing to test the product, Chowdhary said, including breeding companies, academic partners and grower associations. The robot has been deployed in corn, sorghum, wheat, soybeans, vegetable crops, orchards and other crops.
Collecting data
Carl Bernacchi, Ph.D., is a research plant physiologist with the U.S. Department of Agriculture’s Agricultural Research Service and an affiliated associated professor of plant biology at the University of Illinois. Bernacchi is also deputy director of EarthSense’s TerraSentia project.
The focus of the program is to bridge high-throughput phenotyping with the ability to measure plant traits at a high rate in a short period, Bernacchi said. “We take these high throughput techniques and use them to better understand the genetic drivers for crop productivity.”
Bernacchi’s research has focused on bioenergy sorghum, but the work “extends well beyond just bioenergy crops into food crops as well,” he said.
As a plant physiologist, Bernacchi measures plants in the field and tries to understand how and why they respond to different environmental conditions in certain ways. His work with TerraSentia aims to help streamline the process and get measurements more quickly and with less money.
“This project is really intended to try to figure out ways of automating a lot of the traditional measurements that we make in order to advance not only breeding, but also to understand and advance our understanding of crop response to the environment,” Bernacchi said.
TerraSentia measures the height of the plant, how wide the plant is at the base and how many plants are in a field. Leaf area index is the most complex measurement the robots take currently, he said. This measurement of the crop canopy quantifies the total area of leaves in the total area of ground.
The measurements the TerraSentia robot takes aren’t groundbreaking in themselves, Bernacchi said. Consider an agricultural field that has anywhere from 1 million to 10 million plants, or a standard breeding operation. “No one has ever been able to measure the width, the height, the leaf area index, and the number of plants in a field that size within a couple of hours.”
Some measurements could be taken with calipers and rulers, he said. If they tried to do a similar analysis as the robots by hand, though, it would take an army to complete the task—and much longer than two hours.
Humans are less precise in their measurements than the technologies that have been incorporated into the robot, Bernacchi added. “So we are taking orders of magnitude less time with quite a bit more precision.”
That’s the real strength of TerraSentia and other types of ground robots and why the scientific research being conducted is groundbreaking.
Future applications
Bernacchi said his team has made an incredible number of measurements in the past several months, at multiple points in the growing season.
“We have been able to collect high-quality data at regular intervals from emergence all the way through harvest,” he said.
Researchers can use that information to understand the genetic basis for how crops respond to different environmental conditions.
“Moving into the future, more capabilities could be incorporated to the ground-based robot that would provide meaningful information with regard to pests or pathogens, which usually start from within the plant canopy,” Bernacchi said.
Chowdhary said future applications for the robots would be adding management actions, such as selective spraying or weeding with mechanical weeders. Robots could do more precise tasks than large agricultural equipment.
“We have active work going on in dexterous arms for the robot so that we can do operations in various polycultures,” he said. That technology would help organic growers, for example. Small growers could benefit by not having to invest in large machines.
Chowdhary said, “We’re really working hard to bring it to the growers, so any help and encouragement we can get from growers is very valuable to us.”
A National Science Foundation Small Business Innovation Research Award provides ongoing funding for EarthSense’s work. To learn more about TerraSentia, visit earthsense.co.
Shauna Rumbaugh can be reached at 620-227-1805 or [email protected].
