postdoctoral research assocoate University of Arizona Lawrenceville, Georgia
Body of Abstract: Computer vision-based systems have freed academic researchers from time-consuming and resource-demanding plant phenotyping tasks. However, most current plant phenotyping systems focus on monitoring morphological and physiological traits. Analysis of plant color traits over time is a unique addition for current plant phenotyping systems. Therefore, we present the open source and affordable plant phenotyping software pipeline, SMART (Speedy Measurement of Arabidopsis Rosette Traits).
Beyond the computation of morphological traits of the Arabidopsis rosette such as leaf area, leaf width/height, and whole-plant mass center location, SMART can compute leaf-specific traits over time as a proxy of physiological function and abiotic stress response. SMART is successful across different application scenarios with varying imaging setups. For example, SMART was able to quantify growth of the Arabidopsis rosette using a top view image captured by a Raspberry Pi camera. SMART also demonstrated its ability to analyze leaf-specific color changes over time when plants experienced changes in nutrient availability using the imaging robot developed in the OPEN Leaf project. Furthermore, SMART can use standard color checkers to calibrate color of affordable consumer cameras.
We believe that leaf-specific color analysis of the Arabidopsis rosette with the SMART phenotyping pipeline will benefit the whole plant science community to discover and characterize previously unidentified phenotypes.