High Throughput Root Phenotyping in Aeroponics: One Step From Seed to Root Growth Parameters

If modern high-throughput phenotyping platforms generate multitude of image sequences at high spatial, spectral and temporal resolutions, unlocking the full information content of these images remains a major challenge. The aeroponics phenotyping platform at UCL has been designed in order to simplify this issue. It combines backlight illumination, which eases considerably the segmentation of roots from their background, and a very fine temporal resolution (two hours), which enables innovative tip tracking algorithms to estimate dynamic traits and more detailed morphological characteristic of root architecture. Growth rates, new root formation, branching sequences and other dynamic root traits can be extracted through specific, model inspired, data analysis pipelines. Multi-object tracking techniques on real images and machine learning techniques trained on simulated root systems are currently tested to reconstruct the entire root system from tip detection. This platform, which offers a single pipeline from the seed to root growth parameters, is part of the EPPN2020 transnational access portfolio.