Determination of some Growth Characteristics of Seedlings in Controlled Environment by Image Processing

Abstract

Determinate the growth of nursery plants allows assessing the success of breeding methods and identifying factors that directly affect the nursery process. However, accurately assessing the growth of nursery plants mainly by recording daily by hand leads to a waste of resources in both man power and time. The paper presents the processing method and results achieved in determining some growth characteristics of seedlings in a controlled environment by image processing. The characteristics of seedlings in membrane houses determined include plant identification, plant height and number of leaves. Aruco marker is used to identify plants, align images and serve as a reference to calculate plant height. Image analysis and color filtering techniques are used to separate plants from the background. The YOLOv8 artificial neural network model has been further trained with labeled leaf images to count the number of leaves.