Robots, 3-D Imagery Helps Identify Disease, Pest Pressure in Crops

Point cloud of peanut field
This is a partially constructed point cloud of a peanut field that, when complete, will allow UGA scientists to determine height, width, leaf cover, growth and disease anomalies for individual plants and to track those plants through the growing season.

The combination of 3-D imagery and autonomous robots may one day identify crop issues earlier and more efficiently, according to UGA scientist Glen Rains.

Detecting the beginning stages of a disease or identifying insect pressure early in the growing season means a farmer can treat and manage what might be a small problem before it becomes out of control. Such issues include lack of growth, nutrient deficiencies and nematode pressure.

“Our idea is to get in there before symptoms are advanced and reduce the impact of the problem,” Rains said. “If you see a problem early and it turns out you have nematodes in certain areas of your field, then you can make a chemical application for nematode control for those areas. Right now, the assumption many farmers make is that if you have nematodes, then you treat the whole field. In this case, if you could pinpoint exactly what the problem is, you could save a lot of money by applying treatments where they need to go.”

During research in 2016, Rains mounted a camera on tractor that took thousands of images as it moved through a field. As the tractor goes back and forth, it captures multiple images at different perspectives. These pictures were put into Georgia-Tech-developed software that reconfigured the 3-D structures of the plants.

Once the images were processed, Rains sent an autonomous robot into the field, where a potential problem was observed. It took leaf and soil samples of the plants showing early signs of disease, nutrient imbalances and insect infestations. Then, a diagnosis was made to identify the problem.

This technology also shows the growing rates of crops with more accuracy, according to Rains.

“If a farmer sees there is a nutrient imbalance, they could do something right away,” Rains said. “A lot of plants go through different stages, and you want to make sure at certain growth stages that you’re optimally providing the nutrients they need for flowering, fruit production or whatever it is. You can make that change immediately.”

Rains’ research will continue this summer. He believes another four or five years are required to work out the details before this technology is used in peanuts. He also believes there is potential to have a ready-to-apply remedy on board the robot.

By Clint Thompson