Smart Technologies in Agri-Horti-Water-Food
Morphological descriptions of ornamental crops are important for several different reasons, e.g.,
registration purposes, and assessment of distinctness, uniformity and stability (DUS testing). Currently,
these descriptions are obtained manually by highly skilled and trained specialists. Automation of this
description process through application of machine learning would lead to more consistent, objective
and high-quality results. The technology could easily be spread among stakeholders all over the world,
which would place the internationally-oriented Dutch floriculture industry in a prime position.
In this project, two species will be considered: Gerbera and Rose. On the basis of expert knowledge and
data availability a subset of characteristics will be selected for which automatic classification will be
implemented on the basis of color images (RGB) of the flowers. The classification will be done with state-
of-the-art machine learning approaches such as deep learning. The obtained characteristics will then be
compared to existing databases to answer questions related, a.o., to DUS testing. Finally, a software
prototype will be built incorporating the trained classifiers which is able, given an image of a Gerbera or
Rose flower, to present the required classification.
Doel van het project
Relatie met missie/motivatie