Pullman, Washington, USA
March 16, 2015
Wheat breeders at Washington State University are sizing up experimental crops from a new perspective: cameras that see far better than the human eye.
Scientists deploy tractor- and cart-mounted multi-spectral cameras to see how new wheat varieties handle challenges like drought, heat and disease. Results will help breeders and growers choose the best varieties.
“For thousands of years, people have been looking at plants in a field and saying, ‘that one grows well,’” said WSU spring wheat breeder Mike Pumphrey. But there’s a lot our eyes can’t see that a new generation of cameras can.
“Now we have sophisticated cameras that tell us with one image: ‘Does that plant have enough groundwater? Is that line of plants using water more efficiently than the next?’ We’ve never been able to do that with our eyes,” said Pumphrey, who works in the College of Agricultural, Human and Natural Resource Sciences’ Department of Crop and Soil Sciences (http://css.wsu.edu/).
Traits through the lens
“This is a tool, like a tractor or even your hand – a tool that helps us collect data,” said Sindhuja Sankaran, an agricultural engineer in the WSU Department of Biological Systems Engineering (http://bsyse.wsu.edu/). She is working with the wheat breeding program to develop the camera technique.
Rolling through wheat fields, her cameras look at phenotypes – the ways plant genes are expressed physically – using visible, infrared and thermal wavelengths, 3D imaging, even sound waves. Proteins, sugars, starches, cellulose and water in plants all reflect light differently. By measuring these spectral signatures, cameras can unlock secrets of plant physiology.
Jianfeng Zhou, a postdoctoral associate in the WSU Department of Biological Systems Engineering, and Honghong Gao, a visiting scholar, test a 3D camera and ultrasonic sensor, measuring the height of wheat at WSU’s Plant Growth Facility. Traditionally, breeders measure with yardsticks. Their mobile camera method could be much faster and more accurate. (Photo by Seth Truscott, WSU CAHNRS)
A water-stressed or damaged wheat plant, explained Sankaran, looks different to an infrared or thermal camera than to the naked eye. For example, chlorophyll – the green pigment in plant cells – converts visible light into energy and reflects infrared light. Stressed plants will absorb less visible light and reflect more infrared than their healthy counterparts.
Sankaran’s challenge is both to analyze the complex data collected by the cameras and to take them to a whole new level.
She and colleague Lav Khot are working with the WSU Office of Research for permission to put phenotype cameras aboard an unmanned aerial vehicle. If approved by the Federal Aviation Administration, the tiny aircraft would hover over wheat fields, snapping photos and creating spectral data about the hundreds of plots of experimental varieties below.
Seeing the big picture
Last summer, WSU researchers held their first large-scale pilot run of the phenotype cameras. A camera, mounted on a Kubota tractor, rolled through wheat fields outside Pullman at 2 miles per hour, snapping three frames of every test-bed of wheat plants.
Vic DeMacon, a scientific assistant with the WSU wheat breeding program, uses a multi-spectrum camera to scan for canopy temperature in wheat plants during Pullman field tests last summer. (Photo by Mike Pumphrey, WSU)
WSU wheat varieties are tested in 80-square-foot plots across 10- and 20-acre fields.
“That’s 6,000 plots in one field,” Pumphrey said. “That’s the kind of scale we’re working on – tens of thousands of lines.”
On that scale, putting a new lens on the big picture could change the game for wheat breeders who want to keep only the best plants and rule out runners-up.
“The real challenge of breeding is to throw away the poor ones and keep the good ones,” Pumphrey said. “When we know that a line is drought tolerant or grows better, that allows us to be more efficient with all our efforts. It gives us greater confidence in those decisions early on.”
A 3D camera image spotlights plant height as a trait, useful in breeding programs. Jianfeng Zhou, a postdoctoral associate with the WSU Department of Biological Systems Engineering, analyzed data and created the image using the Matlab program. (Image by Jianfeng Zhou, WSU)
“It makes things easier and faster,” said WSU winter wheat breeder Arron Carter. “It may only take five minutes to fly over my plot ground, where it would take me one or two hours to walk through it to take notes.”
While helping breeders quickly identify plants with higher yield potential, the technology could also carry over to commercial farming for use in fertility, water, disease and pesticide management.
“This is the future of agriculture,” Carter said.
Topics
Species
