Rice researchers at the Texas A&M AgriLife Research Center at Beaumont have begun a project that utilizes unmanned aerial vehicle, UAV, data to speed up rice cultivar selection and breeding.
The team will use UAVs to capture real-time images of rice crops, extract crop phenotypic traits from the images, and ultimately analyze that information to discover superior, high-yielding rice genotypes.
Key research objectives include:
- Quantify key phenotypic traits of rice growth and development.
- Capture UAV images of rice genotypes at key rice growth stages.
- Develop advanced image processing algorithms to extract key phenological, morphological and architectural traits for key rice growth stages.
- Develop a digital rice selection system that screens for best-performing genotypes through data integration and multi-trait decision making.
Yubin Yang, senior biosystems analyst at the Beaumont center, is directing the Texas A&M AgriLife Research project, which is funded by a three-year, $650,000 grant from the U.S. Department of Agriculture-National Institute of Food and Agriculture. The project seeks to bypass a major hurdle in data gathering – the labor-intensive and time-consuming process of manual field data collection using skilled labor.
Advantages of UAV technology
“Traditional manual measurement of rice phenotypic traits is very, very time consuming,” Yang said. “It’s becoming more and more challenging to hire qualified experienced staff. UAV technology and advanced image processing could potentially provide a cost effective and reliable alternative. We can use UAVs to capture rice images at key growth stages and develop algorithms to extract different phenotypic traits for hundreds or even thousands of rice genotypes.”
Thousands of UAV images will be gathered as well as ground truth data throughout the rice crop season, Yang said. Multiple UAV flights will be carried out to capture rice images with different camera angles to help analyze stand establishment and gaps between plants.
“A considerable amount of data will need to be integrated and analyzed,” Yang said. “This is the first year of the project and is a learning process for us. Timely capture of UAV images for early rice growth has been challenging due to the small size of the rice seedlings and windy weather conditions. There’s a limited window when you can fly.”
Developing advanced image processing algorithms
While acquiring the UAV data, the team will also develop machine-learning algorithms to identify key traits and select best-performing rice genotypes.
The research project will evaluate key phenotypic traits for breeding selection, such as stand establishment, biomass growth, phenological development and final grain yield.
“We will be developing automated algorithms that can extract phenotypic traits from UAV images taken at critical rice stages, including seedling, tillering, flowering, grain filling and maturity,” Yang said. “The digital rice selection system will be developed through integration of multiple traits to identify best performing genotypes.”
Dou said another potential aspect of using UAV technology will be monitoring plant growth for nitrogen management and disease detection.
“We have an ongoing project to assess rice nitrogen status from UAV images,” Dou said. “UAV-based diagnosis of plant nutrient and other stresses has tremendous potential, especially for rice with limited field access due to a flooded production system.”
“This proposed project represents a major effort in delivering an integrated UAV imagery-based decision-making system to rice breeders and researchers,” Yang said. “It will be an indispensable tool to greatly improve rice breeding and phenotyping efficiency.”
Joining Yang on the project team are fellow AgriLife Research scientists Stanley Omar Samonte, hybrid rice breeder; Fugen Dou, crop nutrient management; Ted Wilson, center director and holder of the Jack B. Wendt Endowed Chair in Rice Breeding; Tanumoy Bera, post-doctoral research associate, all in the Texas A&M Department of Soil and Crop Sciences; and Jing Zhang, associate professor in computer science, Lamar University, Beaumont.