In-field crop row stereo-reconstruction for plant phenotyping - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

In-field crop row stereo-reconstruction for plant phenotyping

Résumé

This article presents a methodology for crop row structure characterization developed for phenotyping purpose. It is based on 3D modeling performed using Structure from Motion. The model provides an accurate crop row reconstruction and serves as a basis to retrieve plant height and leaf area. To do so, a robust discrimination method using both color and height information is proposed to separate vegetation from soil. The point cloud is scaled and the plant surface is approximated by a triangular mesh. The efficiency of our method was assessed with two data sets collected in outdoor conditions. Height estimation was extremely accurate since both average error and error of reference measurement had the same order of magnitude. Strong correlations and low errors were also obtained for leaf area estimation. The proposed method thus provides an interesting tool for phenotyping.
Fichier principal
Vignette du fichier
mo2014-pub00041140.pdf (1.19 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01012196 , version 1 (25-06-2014)

Identifiants

Citer

S. Jay, Gilles Rabatel, N. Gorretta. In-field crop row stereo-reconstruction for plant phenotyping. Second International Conference on Robotics and associated High-Technologies and Equipment for Agriculture and forestry (RHEA-2014), May 2014, Madrid, Spain. 10 p. ⟨hal-01012196⟩
234 Consultations
295 Téléchargements

Partager

Gmail Facebook X LinkedIn More