Growing Technology

Weilong He loves the challenge of solving a problem. So when He had the opportunity to improve the process of tracking crop growth for plant breeders, he jumped at the chance. 

Traditionally, plant breeders have had to plod through densely packed rows of crops on foot to manually measure and observe the size, quality and angle of every leaf. This tedious work can come at an even greater cost if they can’t identify problems fast enough to ensure the best yield. 

To help, He, a graduate research assistant in the Automation and Robotics Lab in the Department of Biological and Agricultural Engineering, has spent the past year at NC State University exploring how to use machine learning in tandem with drone images to assess crop leaves faster. 

“I like that you work on the problem many times and try almost every solution until you find the best way to solve the problem,” says He, who earned his undergraduate degree in electronic information technology from Wuhan University of Technology in China. 

Part of the inaugural cohort of the Norma L. Trolinder N.C. Plant Sciences Initiative Graduate Student Endowment Award, He, who is pursuing his Ph.D. in agricultural engineering, feels his work is the next chapter of research started by endowment benefactors Norma and Linda Trolinder. The mother-daughter duo co-founded South Plains Biotechnologies Inc. in the 1990s and focused their work on plant genetics. Linda Trolinder went on to hold research and development roles with companies such as Bayer Crop Science and BASF. 

“The Trolinder award has enabled me to devote more time and resources to my research in plant sciences,” He says. “It has facilitated my participation in academic conferences and field trips, providing invaluable opportunities to engage with esteemed scholars in plant science, which have significantly enriched my understanding and contributed to the progress of my research endeavors.” 

One of those endeavors is how to better automate the process of measuring the angles of corn leaves. Breeding specialists currently have to do it by hand, which can take many hours to complete. The angle is important for determining the light and water interception within the tightly packed rows of crops. 

To save breeders time, He is developing a machine-learning algorithm to generate measurements for every leaf based on drone images of a corn field. The technology, so far, is promising with an accuracy rate of more than 90%. He and his advisor, Lirong Xiang, assistant professor of biological and agricultural engineering, have been working with plant breeders, including NC State faculty Joseph Gage and Rubén Rellán Álvarez, as well as Iowa State University researcher Patrick Schnable, on leveraging this technique for large-scale studies in breeding programs. 

Another project centers on developing an algorithm to identify healthy, diseased and withered leaves on tomato plants. The long-term goal, He says, is to have the algorithm correctly label the health of the leaves and then eventually predict the health of the plant. 

Xiang, who runs the Automation and Robotics Lab, says the goal of this kind of research is to develop automated solutions for agriculture by leveraging technologies such as robotics, AI and machine vision. 

“We aim to develop fully automated systems for collecting phenotypic traits, such as leaf angle and disease severity, to save researchers time and costs associated with manual measurement, thereby ultimately accelerating breeding programs,” Xiang says. 

And watching He make meaningful progress in this work, says Xiang, is rewarding. 

“I chose to do research in agriculture because I really like plants. Now, I’ve combined my love of electronic and computer engineering with agriculture.”
– Weilong He

“He is someone who stays focused and is self-motivated, consistently demonstrating a proactive approach to learning and problem-solving,” she says. 

He is also developing a versatile skill set by helping build devices such as a remote-controlled spore trap system for sampling cucumber fields and an automated pruning robot for strawberries. 

The best part of his work, though, is blending his passion for the technical world with that of the natural world. 

“I chose to do research in agriculture because I really like plants,” He says. “From a very small age, I planted a lot of flowers and other plants in my own garden. So, when I graduated with my bachelor’s degree, I wanted to do research about plants. Now, I’ve combined my love of electronic and computer engineering with agriculture.” 

---

All photos by Amanda Kerr.