Phenotyping and Plant Health
By Thomas Skernivitz, Senior Editor, American Fruit Grower and American Vegetable Grower
Take Purdue University’s nationally ranked college of agriculture and its equally heralded engineering school, mix them together, and what does one get?
The No. 1 agricultural and biological engineering (ABE) program in the country
The dream job of Associate Professor Jian Jin
A breakthrough plant health tool that growers can hold in their hands
All of the above
The answer, of course, is “D” — as in In-DEE-ana … West Lafayette to be specific.
“Our Department of Agriculture and Biological Engineering is right in-between, at the overlapped part of these two top-10 colleges of ag and engineering,” Jin says. “That’s probably why we're No. 1, but there are a lot of other reasons.”
While that population amounts to predominantly corn and soybean farmers, Jin is quick to note that he and his department’s anchor technology — plant phenotyping, or plant sensing — can and should apply to growers of all types across the country, specialty crops included.
Foremost among those, Jin says, is a faculty-grower relationship that remains a priority regardless of his department’s ever-increasing prestige.
“We have a very, very talented faculty. We are working hard and always thinking about the farmers, the customers,” Jin says. “For example, we work closely with our Extension network and keep thinking about the farmers — the countryside — when we do our research. After years and years of work, we’ve gotten our reputation and a lot of trust from the whole state of Indiana.”
“Most U.S. farmers are still relying on the same way their grandparents determined plant health, which is looking at the plants with their own eyes and guessing,” Jin says. “We feel ashamed that we still haven't helped the farmers to make better decisions. But I'm seeing the light at the end of the tunnel. With more progress, we should be able to change this game and allow farmers to make wiser decisions very soon.”
Innovation from Day One
Plant phenotyping — defined as the process of measuring and analyzing observable plant characteristics — took flight at the university in 2013 as part of a new plant science initiative. While the resulting research has always involved basic sciences, such as engineering and breeding, the then-relatively new concept of phenotyping was the focal point from the get-go.
“Under the great leadership of Karen Plaut [currently Purdue’s Executive Vice President for Research] and others, we started to hire faculties, and we invested a lot of resources to build the phenotyping facilities to highlight Purdue as one of the leading universities in the nation in plant phenotyping,” he says.
One of those hires, in 2015, was Jin himself. He had spent the previous five years as a Research Scientist and Technology Leader with DuPont Pioneer (now Corteva Agriscience) in Johnston, IA. Before that, he earned his doctoral and post-doctoral degrees at Iowa State University, which currently boasts the nation’s No. 2-ranked ABE program behind Purdue.
“Not only the so-called larger-scale automatic, high-throughput phenotyping facilities, including our greenhouse facilities and field facilities, but also we try to miniaturize some of the technology into small handheld devices and put the technologies into ordinary people's hands. Not just university professors and industry collaborators but ordinary farmers and greenhouse growers.”
Since then, Jin has made the most of his Purdue career, starting with his founding of the ABE Plant Sensor Lab soon after his arrival and validated by his department’s current No. 1 national ranking.
“With this lab, we keep innovating,” Jin says. “Not only the so-called larger-scale automatic, high-throughput phenotyping facilities, including our greenhouse facilities and field facilities, but also we try to miniaturize some of the technology into small handheld devices and put the technologies into ordinary people's hands. Not just university professors and industry collaborators but ordinary farmers and greenhouse growers.”
Handheld Hyperspectral Imaging
Having established himself as a professor and researcher, Jin turned entrepreneur in 2018 when he launched an ag-tech startup out of the ABE Plant Sensor Lab. The namesake product of LeafSpec LLC — patented by the Purdue Innovates Office of Technology Commercialization — is billed as the first (and still only) portable hyperspectral crop leaf imager of its kind.
The handheld device and software provide accurate and early detection of plant diseases, nutrient deficiencies, and stresses from chemical sprays. This can be done within seconds, without damage to the plant, according to Jin, the company’s CEO and President. The device then relays the measurement results to the farmer’s or scientist’s smartphone.
“This information can guide users to make quicker and smarter decisions on fertilizing, spraying, and irrigation, saving their cost and improving the yield,” Jin says. “The result is improved food security: additional crop resiliency and production that leads to more crops available for harvest and worldwide consumption.”
LeafSpec markets multiple models of its device to cover different crops, such as corn, soybeans, wheat, rice, and grapes. The company, which sold its first device in 2020, this summer reported sales of nearly $400,000. Most of LeafSpec’s current users, according to Jin, are plant scientists and breeders from universities and large companies, such as Bayer, FMC, Michigan State University, and Purdue.
Precision Ag Endgame
Jin, 42, has experienced the benefits of phenotyping from two perspectives during his career. He grew into the technology while working in the breeding industry for DuPont Pioneer.
“Breeding is basically a game of selection. In your hands you have a lot of seeds, and you’ve got to choose the best genotype or the best traits to provide farmers the best seeds so they can grow higher-quality products,” Jin says. “But how can we more quickly select the best seeds? That's the question.”
Phenotyping provided the answers by more efficiently measuring and detailing seed performance. In turn, Jin was able to more accurately determine which seeds posted the top yields and fit best in specific environments while also offering improved resistance to stress, such as drought and disease.
“In order to accomplish precision agriculture, you have to have three components,” he says. “You need to have the sensor to measure what is going on. Then the computer to make decisions. And then the computer will command the actuators to apply treatments in the field.”
“Phenotyping in breeding helped us to deliver better seeds to the farmers’ hands earlier,” he says.
At Purdue, Jin applies phenotyping to a plant’s leaves rather than its seeds. In turn, he is trying to “close the loop” that is precision agriculture.
Comparing those three components, the actuators and computers are “very mature,” Jin says. Yet precision agriculture has not reached its potential across the world, he adds, because high-quality sensors are still lacking.
“That is why corn growers, when their plants are at the V6 stage, are still wondering how much nitrogen they should put into the field. The less they put on, they may risk their yield. The more they put on, that costs them more and will pollute the environment because the nitrogen is going to flow into their local water system. It's a very hard decision to make.”
Room for Improvement
Previous crop remote sensing platforms never lived up to their billing, Jin says. He saw that during his first two years at Purdue while developing large-scale phenotyping facilities.
“I flew a lot of drones with the cameras, looking at the field and taking images, processing the images, and trying to give the answers,” he says. “But then I learned a lot about why, after so many years of the development of the sensors, farmers were still not buying those.”
Jin blames “noisy” practical environmental issues for past failures. For example, when taking images from a drone, the primary lighting source — the sun — is constantly changing in angle, intensity, and color. The clouds are moving and leaves shaking in the wind. The top and bottom leaves of a plant are completely different while they’re mixed in the images.
“All those noises are combining toward the data set. That was why it was not working for the corn growers,” Jin says. “We wanted to change this.”
LeafSpec has analyzed and eliminated most of those “major noise issues” while transferring the hardware and software technologies of big hyperspectral imaging stations to a small handheld device.
“When the noises are gone, the signal shows up and we can provide better quality. With better and better quality, we are seeing the day when we can enable precision agriculture,” Jin says. “We have 800 million farmers in the world. We want to benefit more people, and we want to make the technology easily accessible at any location.”
What About Specialty Crops?
Despite being headquartered in the heart of row crop country, LeafSpec does cater to fruit and vegetable growers domestically and internationally. One of its customers is Farm2U, a vertical company that provides services to hundreds of apple growers in the Kashmir region of northwest India.
“That’s a world-famous valley for apple production where they used LeafSpec locally to scan the leaves,” Jin says. “They actually sponsored another research project at Purdue for us to help them process their images, and the results showed that, with the LeafSpec images, we were able to identify, yes or no, is there a disease? And if there is any disease, we know which type of disease. That helps farmers a lot.”
Early detection is vital in these cases, Jin says, and he believes hyperspectral imaging technology can be a significant tool to help growers.
On a related note, Jin relays the story of the devastating 2022 outbreak of fire blight in New York state apple orchards.
“Fire blight is one of the many diseases of that type that, when you notice something is going on in your orchard, it's already too late. There is nothing you can do. You can only wait and watch,” Jin says. “Very few insurance companies would like to provide a service for this kind of coverage, or their charge is very, very high. So, the orchard growers are really depending on luck. It's a high-risk business.”
Early detection is vital in these cases, Jin says, and he believes hyperspectral imaging technology can be a significant tool to help growers. “We are combining high quality, hyperspectral imaging, and high spatial resolution. We are proud that we have introduced and innovated that from Purdue.”
Growers are starting to take notice. Jin attracted attention from local fruit growers, vineyardists, and greenhouse vegetable growers while conducting a research seminar at Cornell University’s AgriTech Section in Geneva, NY. He eventually received an invitation to compete in November as one of 20 finalists in Grow-NY, an annual competition designed to enhance food, beverage, and agricultural innovation in upstate New York.
LeafSpec, one of just three Midwest companies in the mix, earned accolades from the Grow-NY committee.
“Generally, the response I heard in New York,” Jin says, “is that they found the coolest sensor technology from the Midwest.”
