Skip to Main content Skip to Navigation
Master Thesis

Développement de méthodes de phénotypage haut débit basées sur l'analyse d'images pour le suivi de traits d'intérêts sur une collection d'ignames (D. alata) en Guadeloupe

Abstract : Yam is a tropical geophyte plant with a lianascent habit, characteristic of great phenotypic importance. It is grown for the starch richness of its tubers and is a staple food in many tropical countries. However, the weaknesses and shortcomings of the improved varieties on the quality traits will lead to the growth of the sector in the West Indies and hamper the adoption of new varieties in West Africa. To accelerate progress in variety selection, it is necessary to increase the precision and throughput of phenotyping studies. In yam, the color of the tuber chair, its oxidation, and the nutritional quality of the yam products, partly through the size and shape of the starch grains. To increase the speed of acquisition of measurements of these characters, we tested three phenotyping methods on image analysis within a working collection of 12 varieties of yams grown in two contrasting soil and climate environments. We maintain the performance and application conditions of these methods. Improvement of pre-existing methods and the design of new tools for segmentation, detection and correction of images are proposed. In order for all the methods tested and properly developed for the technical and financial constraints of tropical features, a trade-off between scientific relevance and feasibility (low cost and viability) was sought and quantified.
Document type :
Master Thesis
Complete list of metadatas

Cited literature [37 references]  Display  Hide  Download

https://dumas.ccsd.cnrs.fr/dumas-02138698
Contributor : Agrocampus Ouest <>
Submitted on : Tuesday, June 4, 2019 - 8:21:34 AM
Last modification on : Monday, July 6, 2020 - 3:38:22 PM

File

2018_Mell_Pauline_APVV_FGA.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : dumas-02138698, version 1

Collections

Citation

Pauline Mell. Développement de méthodes de phénotypage haut débit basées sur l'analyse d'images pour le suivi de traits d'intérêts sur une collection d'ignames (D. alata) en Guadeloupe. Sciences du Vivant [q-bio]. 2018. ⟨dumas-02138698⟩

Share

Metrics

Record views

36

Files downloads

53