PubAg

Main content area

Potential use of seed morpho-colourimetric analysis for Sardinian apple cultivar characterisation

Author:
Sau, Silvia, Ucchesu, Mariano, D'hallewin, Guy, Bacchetta, Gianluigi
Source:
Computers and electronics in agriculture 2019
ISSN:
0168-1699
Subject:
Malus domestica, apples, breeding programs, color, computer software, cultivars, databases, discriminant analysis, early selection, fruits, image analysis, phenotype, scanners, seeds, trade
Abstract:
Linear discriminant analysis (LDA) of morpho-colourimetric features, obtained through scanned images of apple (Malus domestica Borhk.) seeds from 25 Sardinian native cultivars, was employed to validate this approach as an easy and fast method to discriminate apple cultivars.Digital images of seeds were acquired using a flatbed scanner that measured 154 morpho-colourimetric features using ImageJ software. The data were used to build a database and statistically processed by LDA in order to identify and investigate phenotypic characteristics of apple cultivars. Afterwards, LDA of the database was used to investigate skin colour differences according to the apple variety colour descriptors established by the International Union for the Protection of New Varieties of Plants (UPOV) and the International Board for Plant Genetic Resources (IBPGR).The results evidence that by LDA it is possible to classify apple cultivars and highlight the presence of three possible synonymy groups. Moreover, when considering skin colour, LDA correctly classified a high percentage of both ground and over colour, as described by the IBPGR; in particular, using LDA it was possible to correctly identify 90.5% and 93.5% of green-yellow and red apple groups, respectively.The results of this investigation prove that image analysis of apple seeds is a valuable approach to study and characterise cultivars, as well as to identify important trade characteristics of fruit, such as skin colour, resulting in a useful tool for early selection of fruit characteristics in breeding programmes.
Agid:
6379489