Figure 1. Overview of the RootXplorer platform. (IMAGE)
Caption
RootXplorer is a computer vision-based phenotyping platform designed for high-throughput 3D imaging and quantification of root penetration traits/phenotypes in multiple plant species. RootXplorer integrates three main components: A) A Phytagel cylinder system (CS) for growing plant roots under artificial soil compaction conditions. Bottom Phytagel layer, BPL; upper Phytagel layer, UPL. B) A root architecture 3D imaging cylinder (RADICYL) system to scan plants in the CS and create 360-degree images or videos of roots in 72 frames. C) An image analysis pipeline (IAP) for automatic image cropping, segmentation and trait extraction. Image segmentation was conducted by deep learning models trained with a U-Net++ convolutional neural network (CNN) architecture. After segmentation, root penetration traits for each cylinder/plant are extracted via computer vision and used to calculate penetration phenotypes. Upper root count, URC; upper root area, URA; bottom root count, BRC; bottom root area, BRA; root count ratio, RCR; root area ratio, RAR. D) Data generated from RootXplorer is then used to assess the impact of mechanical impedance on root system growth, analyze root penetration dynamics, and evaluate phenotypic variation in root system penetrability across diverse plant species and accessions.
Credit
The authors
Usage Restrictions
Credit must be given to the creator.
License
CC BY