Accepted_test

Development of an algorithm for separating densely contiguous wheat grains in 2D images

by Dmitry R. Avzalov | Evgeniy G. Komyshev | Dmitry A. Afonnikov | Novosibirsk State University, Novosibirsk, Russia | Institute of Cytology and Genetics, SB RAS, Novosibirsk, Russia | Institute of Cytology and Genetics, SB RAS, Novosibirsk, Russia; Kurchatov Genomic Center of the Institute of Cytology and Genetics, SB RAS, Novosibirsk, Russia

Abstract ID: 512

Event: BGRS-abstracts

Sections: [Sym 6] Section “Analysis of plant images for solving phenotyping problems”

Counting grains manually is a labor-intensive and time-consuming task. Development of approaches based on image analysis by computer vision methods is relevant nowadays due to low cost and ease of use. Grains in such images are located close to each other, which complicates the process of their recognition.

The algorithm consists of 2 main steps: preprocessing the image to obtain contours of grain groups and analyzing the obtained contours to separate contiguous seeds. The preprocessing consists of reducing the resolution of the image, removing noise using the Gaussian filtering method, and obtaining a binary mask using the Mean shift clustering algorithm and the OTSU algorithm. A combination of the corner point detection algorithm and the elliptical contour analysis algorithm was used to analyze the obtained contours.

The metric CR (correct ratio) was used to evaluate the performance of the algorithm. The average value of the metric was 96%, which is a higher result compared to other classical computer vision approaches.