Digital image analysis to assess flower productivity and foliage retention in garden roses throughout the growing season
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
2019 International Society for Horticultural Science. All Rights Reserved. A good ornamental garden rose needs to maintain good foliage coverage and abundant flowers throughout the growing season. Unfortunately an evaluator or breeder generally only has the time to assess these characteristics 2 or 3 times during the 6 to 9 month growing season. Consequently much is missed and the best plants may not be identified. Thus a rapid phenotyping procedure which can be done on a weekly basis is needed. Overhead pictures of a series of rose plants were taken on a weekly or monthly basis during the growing season using a Samsung Galaxy Tab4 which was mounted on a selfie stick. At the same time, these plants were assessed for their percent defoliation and flower intensity. Algorithms were developed to estimate the rose canopy cover, density and the percentage of flower cover based on image segmentation and clustering in red-green-blue (RGB) and hue-saturation-value (HSV) color spaces. These algorithms were able to distinguish the flowers irrespective of their color from the leaves. The analysis was capable of calculating the percent of the canopy covered with flowers but not the percent defoliation. This approach, incorporating an unmanned aircraft system (UAS) or ground-based system equipped with cameras and an analytical pipeline, has potential of rapidly acquiring plant development information essential for the proper evaluation of garden roses as well as for field physiology research.
