Outdoor Plant Segmentation With Deep Learning For High-throughput Field Phenotyping On A Diverse Wheat Dataset

This article discusses the use of deep learning approaches for robust and automated segmentation of leaves and other backgrounds in high-throughput field phenotyping. The study presents a workflow based on DeepLab v3+ and a diverse annotated dataset of 190 RGB images. Results show that a small but carefully chosen and annotated set of images can provide a good basis for a powerful segmentation pipeline, with potential for further development in the future.