Skip to Main content Skip to Navigation
Conference papers

High throughput field phenotyping (HTFP) of wheat and weed cover in field experiments using RGB images: assessment of crop-weed competition with a simple ecophysiological model

Abstract : The early stages of growth for two winter wheat cultivars, Apache and Rubisko, were studied in field experiments based on destructive measurements and visible images. They cover the period from the three-leaf stage to tillering at four sampling dates. Maps of fractional vegetation cover (FVC) were established for both the crops and weeds. FVC was automatically determined from the images with an SVM-RBF classifier, using Bag of Visual Words vectors as inputs. The heterogeneity in populations and crop-weed competition were studied using descriptive and inferential statistics. The impact of weeds on crops was evaluated by comparing the results with simulations under unstressed conditions.
Complete list of metadata

https://hal-agrosup-dijon.archives-ouvertes.fr/hal-03335965
Contributor : Admin Agrosupdijon <>
Submitted on : Monday, September 6, 2021 - 4:26:36 PM
Last modification on : Tuesday, September 14, 2021 - 2:44:01 PM

Identifiers

  • HAL Id : hal-03335965, version 1

Citation

Christelle Gée, Victor Mignon, Laurence Dujourdy, Emmanuel Denimal. High throughput field phenotyping (HTFP) of wheat and weed cover in field experiments using RGB images: assessment of crop-weed competition with a simple ecophysiological model. European Conference on Precision Agriculture (ECPA2021), Jul 2021, Budapest, Hungary. ⟨hal-03335965⟩

Share

Metrics

Record views

28