Skip to main content
ARS Home » Midwest Area » Madison, Wisconsin » U.S. Dairy Forage Research Center » Dairy Forage Research » Research » Publications at this Location » Publication #357791

Research Project: Redesigning Forage Genetics, Management, and Harvesting for Efficiency, Profit, and Sustainability in Dairy and Bioenergy Production Systems

Location: Dairy Forage Research

Title: Using machine learning to identify and analyze red clover space plants in breeding nurseries

item Riday, Heathcliffe
item Sindic, Caleb

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 11/1/2018
Publication Date: 3/24/2019
Citation: Riday, H., Sindic, C.M. 2019. Using machine learning to identify and analyze red clover space plants in breeding nurseries. International Forage and Turf Breeding Conference, Lake Buena Vista, Florida. 24 to 27 March 2019.

Interpretive Summary:

Technical Abstract: We have developed a program capable of identifying individual space-planted red clover breeding nursery plants in digital images. The program takes advantages of pre-trained neural networks and other machine learning methods. The program is very robust and can identify red clover plants with a 99% accuracy, even when plants of interest are in close proximity to each other, overlap slightly, or are surrounded by complex vegetation, such as turf grass and weeds. The correct within plot plant position was determined 93% of the time. Images fed into the program do not require the camera used to take the images to be in exactly the same position in relationship to the red clover plants from image to image. For identified plants, the program can count the number of leaves on the plant. This plant leaf count value was found to be highly correlated with field based human visual ratings (a proxy for biomass yield) of these same plants with an r2 of over 70%. Future work will improve the program’s accuracy and expand the number of traits phenotyped.