Assessment of flower colour of tuberous begonia using image analysis

Project Details


Main research question/goal
Can we develop a tool with image analysis to objectively determine the colour of flowers? The assessment of flower colour is very difficult. Usually it is done (visually) by experts using a colour chart. In the context of breeding activities or for DUS and VCU research (criteria for assessing new cultivars) a more objective, automatic and non-expert-dependent method for colour assessment is a scientific step forward. Moreover, this method allows for simple statistical processing. This project explores the technical boundaries of image analysis in plant research.

Research approach
Images of flowers and petals of different varieties of tuberous begonias and offspring of crosses in azalea are taken. Flower colour is assessed both by experts in the traditional way (RHS color charts) and by image analysis. Both methods are statistically compared.

By using image analysis for shape and colour determination, quantitative data are obtained that are more related to genetic parameters. This results in more efficient and targeted breeding for azalea (Azalea Innovation Fund). In addition, new varieties (eg tuberous begonia) can be assessed in a more objective way for DUS and VCU research.
Effective start/end date4/05/0931/12/19