Plant biologists welcome their robot overlords

Thursday, February 16, 2017

Old-school areas of plant biology are getting tech upgrades that herald more detailed, faster data collection.

by Heidi Ledford

25 January 2017
A robot measures the crops in an agricultural field near Columbia, Missouri (credit: DeSouza/Fritschi/Shafiekhani/Suhas/University of Missouri)
As a postdoc, plant biologist Christopher Topp was not satisfied with the usual way of studying root development: growing plants on agar dishes and placing them on flatbed scanners to measure root lengths and angles. Instead, he would periodically stuff his car with plants in pots dripping with water and drive more than 600 kilometres from North Carolina to Georgia to image his specimens in 3D, using an X-ray machine in a physics lab.

Five years later, the idea of using detailed imaging to study plant form and function has caught on. The use of drones and robots is also on the rise as researchers pursue the ‘quantified plant’ — one in which each trait has been carefully and precisely measured from nearly every angle, from the length of its root hairs to the volatile chemicals it emits under duress. Such traits are known as an organism’s phenotype, and researchers are looking for faster and more comprehensive ways of characterizing it.

From 10 to 14 February, scientists will gather in Tucson, Arizona, to compare their methods. Some will describe drones that buzz over research plots armed with hi-tech cameras; others will discuss robots that lumber through fields bearing equipment to log each plant’s growth.

The hope is that such efforts will speed up plant breeding and basic research, uncovering new aspects of plant physiology that can determine whether a plant will thrive in the field. “Phenotype is infinite,” says Topp, who now works at the Donald Danforth Plant Science Center in St Louis, Missouri. “The best we can do is capture an aspect of it — and we want to capture the most comprehensive aspect we can.”

The plummeting cost of DNA sequencing has made it much easier to find genes, but working out what they do remains a challenge, says plant biologist Ulrich Schurr of the Jülich Research Centre in Germany. “It is very easy now to sequence a lot of stuff,” he says. “But what was not developed with the same kind of speed was the analysis of the structure and function of plants.”

Plant breeders are also looking beyond the traits they used to focus on — such as yield and plant height — for faster ways to improve crops. “Those traits are useful but not enough,” says Gustavo Lobos, an ecophysiologist at the University of Talca in Chile. “To cope with what is happening with climate change and food security, some breeders want to be more efficient.” Researchers aiming to boost drought tolerance, for example, might look at detailed features of a plant’s root system, or at the arrangement of its leaves.

False-colour images of a bean-breeding trial captured by a camera mounted on a drone (credit: Lav R. Khot/Washington State University & Phillip N Miklas/USDA-ARS)

A need for speed

The needs of these researchers have bred an expanding crop of phenotyping facilities and projects. In 2015, the US Department of Energy announced a US$34-million project to generate the robotics, sensors and methods needed to characterize sorghum, a biofuel crop. Last year, the European Union launched a project to create a pan-European network of phenotyping facilities. And academic networks have sprung up around the globe as plant researchers attempt to standardize approaches and data analyses.

Large-scale phenotyping has long been used in industry, but was too expensive for academic researchers, says Fiona Goggin, who studies plant–insect interactions at the University of Arkansas in Fayetteville. Now, the falling prices of cameras and drones, as well as the rise of the ‘maker’ movement that focuses on homemade apparatus, are enticing more academics to enter the field, she says.

At Washington State University in Pullman, biological engineer Sindhuja Sankaran’s lab is preparing to deploy drones carrying lidar, the laser equivalent of radar. The system will scan agricultural fields to gather data on plant height and the density of leaves and branches. Sankaran also uses sensors to measure the volatile chemicals that plants give off, particularly when they are under attack from insects or disease. She hopes eventually to mount the sensors on robots.

A drone loaded with thermal imaging equipment flies over grapevines (credit: Lav R. Khot/Washington State University)

Sankaran’s mechanical minions return from their field season with hundreds of gigabytes of raw data, and analysing the results keeps her team glued to computers for the better part of a year, she says. Many researchers do not realize the effort and computing savvy it takes to pick through piles of such data, says Edgar Spalding, a plant biologist at the University of Wisconsin–Madison. “The pheno­typing community has rushed off to collect data and the computing is an afterthought.”

Standardizing the technology is another barrier, says Nathan Springer, a geneticist at the University of Minnesota in St Paul. The lack of equipment everyone can use means that some researchers have to rely on slower data-collection methods. Springer has been working with 45 research groups to characterize 1,000 varieties of maize (corn) grown in 20 different environments across the United States and Canada. The project has relied heavily on hand measurements rather than on drones and robots, he says.

Topp now has his own machine to collect computed tomography (CT) images, but processing samples is still a little slow for his liking. He speaks with reverence of a facility at the University of Nottingham, UK, that speeds up its scans by using robots to feed the plants through the CT machine. But he’s pleased that he no longer has to haul his soggy cargo across three states to take measurements. “It’s just endless, the number of possibilities.”

Nature 541, 445–446 (26 January 2017) | doi:10.1038/541445a

Plant Team: Missouri Transect Vinobot in the News

Photo courtesy of Gui Desouza

A growing population gets a robotic assist

Technology aids researchers in examining efficiency, yields of crops

Columbia Daily Tribune

Monday, April 3, 2017

By Megan Favignano

Link to story online:

As soon as University of Missouri researchers plant in the Bradford Research field later this spring, a robot will begin roaming each row and measuring plant features as crops grow.

Ali Shafiekhani, a graduate research assistant, has worked on the project for three years.

He is hoping to further improve the robot, named Vinobot, during his research this summer. Last summer, the robot examined a field with multiple types of corn plants. The robot collects data then sends it to a tall thin pole, or tower with cameras called the Vinoculer, that also is located in the field. Shafiekhani uses the data collected to create 3D models of the plants.

“Sometimes robots can collect more accurate data than from humans,” he said. “We are interested to collect data from the entire growing period.”

Shafiekhani said the robot has three sets of sensors and an arm, both of which are used for collecting data and measuring humidity, temperature and light intensity. Collecting those measurements is called plant phenotyping and it assesses growth and plant yield. Researchers can then see what plant features are common among plants that are more tolerant of drought conditions or have higher yields.

“For this project, our goal is to have systems that autonomously can collect data from the plant and extract features that are helpful,” Shafiekhani said.

Shafiekhani’s adviser, Gui DeSouza, an associate professor of electrical engineering and computer science, and he have collaborated with Felix Fritschi, an associate professor of plant sciences, and Todd Mockler, a principal investigator with the Danforth Plant Science Center. DeSouza said climate change and a growing population worldwide have caused a need to perfect the efficiency of growing crops and the yield of those plants.

The project was funded through a $20 million grant the university received from the National Science Foundation in 2014.

The research team recently published a paper on the project’s progress: “Vinobot and Vinoculer: Two Robotic Platforms for High-Throughput Field Phenotyping.”

The paper documented the system’s accuracy by comparing the measurements or data the robot collected with data collected manually by a research team. While the robot’s data is accurate, Shafiekhani said there is always room for improvement. He’s adding Wi-Fi capabilities to the robot to use during this summer’s research so he can download data remotely instead of going to the field. He’s also working to improve the robot’s navigational system and to enhance the detail of the 3D images created.

DeSouza said there’s no real end for the project. On the engineering side, he said researchers will continue to make adjustments that produce better images and 3D models. And on the plant side, he said researchers will constantly try to improve the crops being grown.

“It’s ongoing. We’re never happy with the end product,” DeSouza said. “We do believe it’s a necessary thing the world requires.”

DeSouza said another team of researchers from engineering and plant sciences has been developing a handheld scanner that collects data as a person walks through the field. That team has studied soybean plants in its research.

mfavignano [at] columbiatribune [dot] com 573-815-1719


Other News Stories covering Dr. DeSouza's Transect Research

Nature:   January 25th 2017

   (large image credited to DeSouza's EPSCoR research)

CoE News:   March 13th 2017


MU News:    March 28th 2017


Stories about this research appeared in 548 international, national, statewide and online media outlets and more than 17.9 million people had the opportunity to view stories online or in print.





EurekaAlert  March 28th 2017


Julie Harker - Brownfield Ag News - interview on March 28th 2017


RFD-TV in Nashville, TN. - interview on March 29th 2017




MU News:    March 30th 2017


NSF NSF “Science 360 News” --  March 30th 2017


KMOV-TV (CBS) in St. Louis

Columbia Tribune


David Sparks’ “” program that runs on 94 radio stations throughout the Northwest!

KOLN-TV in Lincoln, Nebraska


KOMU-TV in Columbia, MO


KTVO-TV (ABC) in Ottumwa, IA


Missouri Ruralist



Tags: Vinobot, Vinocular, Missouri Transect, Missouri EPSCoR, NSF, robotics, autonomous, semi-autonomous, ground vehicle, AGV