Kansas State University using drones to develop better yielding wheat varieties

Kansas State University graduate student Daljit Singh leads the high-throughput phenotyping efforts in India for the Feed the Future Innovation Lab for Applied Wheat Genomics

For the past three years, Jesse Poland, an associate professor of plant pathology at Kansas State University, and a team of researchers have worked in five countries to stay on top of wheat improvement and the most promising technology to help … drones.

“Field seasons in India, Pakistan, and Bangladesh are offset from the field seasons here in Kansas,” said Poland, who is also director of the university’s Feed the Future Innovation Lab for Applied Wheat Genomics. “We can actually be doing work in India during the winter here, learning things and improving the systems, then bring that back and add another season of innovation and improvement here in Kansas. It greatly increases the speed of innovation and testing in our field research.

And the team is using unmanned aerial vehicles (UAVs), or drones, to scout agricultural fields for important data. UAVs can do the work in a fraction of the time that it would take humans.

“What we are trying to do is take these UAVs to the breeding fields to efficiently and quickly measure plant traits,” said Daljit Singh, a Kansas State University graduate student from India who works in Poland’s lab.

Armed with sophisticated, multi-spectral cameras measuring only a few inches, the drones work up and down rows of lush wheat fields, measuring traits such as the plant’s height and vegetation index, or “green-ness” of the plant, which is determined by the amount of light it reflects. The process is known as high-throughput phenotyping because it collects large amounts of information about the plant’s traits, or phenotype.

“The cameras capture the near-infrared light – red, white, green, and blue,” Singh said. “From that, we create a vegetation index because the light reflected from the plant leaves can be associated with the stress levels. The amount of near-infrared reflectance is an indication of whether the plant is going through some type of stress. It can be sick, or have other diseases. It gives us a rapid measurement of what is going on.”

For more complex phenotypes, Poland added:

“It’s similar to facial recognition algorithms; it is using artificial intelligence directly on images. We are developing partnerships with many groups to apply these same approaches to pictures of wheat plants. So, using big datasets of hundreds of thousands of images, you train these algorithms to look at different pictures of wheat plants and identify traits of interest.”

The use of drones and the quickness with which they collect information helps researchers develop massive data sets. Poland says scientists now have the ability to compare genetic differences between thousands of candidate wheat varieties to make better yielding, more heat tolerant, and more disease resistant varieties.

“If you can do more rapid and more accurate selections, then you’re really more efficient in finding that needle in the haystack, the one out of 1,000 that is better all around that can be a new high-performing wheat variety,” Poland said.