Skip to Main content Skip to Navigation
Master Thesis

Mise en place et développement d'un outil de phénotypage assisté par vision artificielle pour la sélection du chou-fleur

Abstract : Despite an increasing number of industrial machines for the characterization of fruits and vegetables, there is no machine today which allows a characterization of an important number of morphological or colorimetric traits of the cauliflower. For these reasons and in order to develop molecular markers, a Vilmorin & Cie (VCO) project assisted by imagery was built in 2012 for the study of 14 traits of the cauliflower. Thanks to a VCO acquisition system, databases and images bases were firstly built. Then, with these bases, image analysis methods and algorithms were developed and statistically validated. Validation methods for the image analysis method and algorithm but also results are illustrated with four morphological traits: diameter and pointed trait for the two dimensional characters, volume and smoothness trait for the three-dimensional characters. Diameter and volume are predicted with regression linear model with a determination coefficient of respectively 98% and 97%. For the pointed trait and the smoothness trait, a classification method by threshold is used to class respectively in two and three categories cauliflower curds, with an error rate of 15% and 34%. This tool has a low throughput but measurement is objective, accurate and reliable (repeatable and reproducible).
Document type :
Master Thesis
Complete list of metadata

Cited literature [48 references]  Display  Hide  Download

https://dumas.ccsd.cnrs.fr/dumas-01082246
Contributor : Agrocampus Ouest <>
Submitted on : Wednesday, November 26, 2014 - 10:02:55 AM
Last modification on : Friday, October 23, 2020 - 4:55:19 PM
Long-term archiving on: : Friday, February 27, 2015 - 10:05:56 AM

Identifiers

  • HAL Id : dumas-01082246, version 1

Collections

Citation

Angéline Bellouard. Mise en place et développement d'un outil de phénotypage assisté par vision artificielle pour la sélection du chou-fleur. Sciences agricoles. 2014. ⟨dumas-01082246⟩

Share

Metrics

Record views

341

Files downloads

152