Skip to main content
Fig. 1 | Plant Methods

Fig. 1

From: Semantic segmentation of plant roots from RGB (mini-) rhizotron images—generalisation potential and false positives of established methods and advanced deep-learning models

Fig. 1

Example minirhizotron images illustrating the varying image qualities and properties used in this study. a and b are images (tree-grassland ecosystem) of the MANIP project dataset; c and d are maize (Zea mays) and olive (Olea europaea) roots from the ATTRACT project dataset, respectively. Both datasets are using different cameras system, focus, resolution, and illumination (see Method section for details on projects/datasets). Subpanels a and b present heterogeneous lighting but overall dark images with a lot of shadows. Both c and d exhibit uniform lighting without significant shadows, but there are distinct variations in soil moisture levels on the outer surface of the MR tubes

Back to article page

Plant Methods

ISSN: 1746-4811