Low-cost vision machine for high-throughput automated monitoring of heterotrophic seedling growth on wet paper support - Archive ouverte HAL Access content directly
Conference Papers Year : 2018

Low-cost vision machine for high-throughput automated monitoring of heterotrophic seedling growth on wet paper support

Abstract

In this communication, we propose a fully automated vision system to monitor the germination and elongation of seedlings positioned in a petri dish. While most existing systems use agar gel as transparent nutritive medium imaged in backlight, we demonstrate that although it provides a reduced contrast, not fully opaque paper can serve as efficient lower-cost medium preventing the well-known problem of seedling joining during elongation. Automatic tracking of elongating seedlings is realized with a minimal pathalgorithm. The three organs (radicle, hypocotyl, and cotyledon) are then segmented. Validation of the accuracy of the system is provided on sugar beet seedling by comparison with the expert-based ground truth.
Not file

Dates and versions

hal-02170866 , version 1 (02-07-2019)

Identifiers

  • HAL Id : hal-02170866 , version 1
  • OKINA : ua17392

Cite

Pejman Rasti, Didier Demilly, Landry Benoit, Etienne Belin, Sylvie Ducournau, et al.. Low-cost vision machine for high-throughput automated monitoring of heterotrophic seedling growth on wet paper support. Computer Vision Problems in Plant Phenotyping (CVPPP 2018), 2018, Newcastle, United Kingdom. ⟨hal-02170866⟩
36 View
0 Download

Share

Gmail Facebook Twitter LinkedIn More