Synapse
⌘+K
Synapse
PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
March 14, 2026

Efficient species segmentation throughout the growing season of oilseed rape-service plant intercropping using transfer learning

View Full Paper
Ask AI
Bookmark
Share

Authors

AJAurélie A. de JongXBXavier BousselinFMFrank de Morsier

Discussion

Loading...

Member takes

Overview

Demonstrates the application of transfer learning to assess multi-species cover in agronomy, highlighting its efficiency.

Key Points

  • To develop a semantic segmentation method using transfer learning for analyzing multi-species covers in intercropping systems.
  • Used 50 images for training a robust semantic segmentation model.
  • Applied transfer learning on the DeepLab convolutional neural network.
  • Conducted a three-year field trial to gather images with various canopy densities.
  • Achieved 96.8% mean accuracy in differentiating vegetation from soil.
  • Developed three-class segmentation models with a maximum accuracy of 96.2%.
  • Successfully identified dynamic competition for light between oilseed rape and service plants throughout all growth stages.

Cite This Study

Jong et al. (2026) studied this question.

synapsesocial.com/papers/69b4fbeab39f7826a300c74ehttps://doi.org/10.1051/ocl/2026002/pdf
View Full Paper
Ask AI
Bookmark
Share