Abstract
Image-based phenotype data with high temporal resolution offers advantages over end-point measurements in plant quantitative genetics experiments, because growth dynamics can be assessed and analysed for genotype-phenotype association. Recently, network-based camera systems have been deployed as customizable, low-cost phenotyping solutions. Here, we implemented a large, automated image-capture system based on distributed computing using 180 networked Raspberry Pi units that could simultaneously monitor 1,800 white clover (Trifolium repens) plants. The camera system proved stable with an average uptime of 96% across all 180 cameras. For analysis of the captured images, we developed the Greenotyper image analysis pipeline. It detected the location of the plants with a bounding box accuracy of 97.98%, and the U-net-based plant segmentation had an intersection over union accuracy of 0.84 and a pixel accuracy of 0.95. We used Greenotyper to analyze a total of 355,027 images, which required 24–36 h. Automated phenotyping using a large number of static cameras and plants thus proved a cost-effective alternative to systems relying on conveyor belts or mobile cameras.
Highlights
Understanding plant genetic effects driving phenotypic differences requires extensive amounts of phenotypic data
For a more reliable detection method, we looked to TensorFlow Object Detection API, designed to locate and find objects in an image and classify them (Huang et al, 2017)
Different configurations of Raspberry Pi (RPi) have been proposed, including overhead fixed cameras, units attached to a camera arm, and units arranged in multi-image octagons to capture three
Summary
Understanding plant genetic effects driving phenotypic differences requires extensive amounts of phenotypic data. Traditional phenotyping approaches are often limited by the time required for data collection and can suffer from batch effects if multiple people are involved in phenotype assessment. Automated phenotyping systems can potentially generate large amounts of unbiased phenotype measurements in a cost-effective manner. A separate internet-connected central computer was connected to this network, centralizing the control of all 180 cameras It received commands via the internet-connection to schedule picture-taking jobs that make all cameras take a picture and transfer it to the central computer. The file name of each image was annotated with a QR code read from the table. The materials and cost of the components used can be found in the system documentation (see link in the Code Availability section)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
