Skip to main content
ARS Home » Pacific West Area » Pullman, Washington » WHGQ » Research » Publications at this Location » Publication #350383
Title: High-throughput phenotyping of seed/seedling evaluation using digital image analysis

Author
item ZHANG, CHONGYUAN - Washington State University
item SI, YONGSHENG - Washington State University
item LAMKEY, RICK - Washington State University
item Boydston, Rick
item Garland-Campbell, Kimberly
item SANKARAN, SINDHUJA - Washington State University

Submitted to: Agronomy
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/26/2018
Publication Date: 5/3/2018
Citation: Zhang, C., Si, Y., Lamkey, R., Boydston, R.A., Garland Campbell, K.A., Sankaran, S. 2018. High-throughput phenotyping of seed/seedling evaluation using digital image analysis. Agronomy. https://doi.org/10.3390/agronomy8050063.
DOI: https://doi.org/10.3390/agronomy8050063

Interpretive Summary: In agronomy and plant breeding, many traits are measured manually using rulers or visual scales with data entered by hand. Image capture and analysis can provide a faster means of quantifying plant growth and health, providing data on multiple traits and seedlings in a single image. This research describes two applications of image capture and analysis; the measurement of coleoptile length in wheat and the measurement of herbicide injury in dry bean. The image capture and analysis was highly correlated with manual measurements of wheat coleoptiles, but using image capture was not a faster way to obtain data on coleoptile growth. The images provided additional data on seedling leaf growth, however. Image capture of herbicide-damaged seedlings was also correlated with the visual assessments, although differences among the rates of the herbicide were not apparent. Image capture and analysis is a useful tool to measure plant growth and health when the procedures for growing the seedlings are adapted to facilitate image analysis.

Technical Abstract: Image-based evaluation of phenotypic traits has been applied to plant architecture, seed characterization, canopy growth/vigor, and root characterization. However, such application of computer vision is not exploited on the measurement of coleoptile length and herbicide injury on seeds. In this study, high-throughput phenotyping using digital image analysis was evaluated on seed/seedling evaluation. Images of seeds or seedlings were acquired using custom-graded digital camera and analyzed using developed image processing algorithms. Results from two case studies demonstrated that it is possible to use image-based high-throughput phenotyping to assess seed/seedlings. In the seedling evaluation study, using color-based detection method, image-based and manual coleoptile length were positively and significantly correlated (P<.0001) with reasonable accuracy (r = 0.69 ~ 0.91), while, using width-and-color-based detection method, image-based and manual coleoptile length was also significantly (P<.0001) correlated (r = 0.89). Improvement of germination protocol designed for imaging device and image processing will increase the throughput and accuracy of coleoptile detection and make image-based seedling measurement more rapid. In regard to the herbicide study, image-processing protocol for extracting color and seed size based features, which could be an indicator to estimate root length and visual symptoms. In presence of the treatment differences, such technique can be applied for non-biased symptom rating.