Temporal dynamics of maize plant growth, water use, and leaf water content using automated high throughput RGB and hyperspectral imaging

Automated collection of large scale plant phenotype datasets using high throughput imaging systems has the potential to alleviate current bottlenecks in data-driven plant breeding and crop improvement. In this study, we demonstrate the characterization of temporal dynamics of plant growth and water use, and leaf water content of two maize genotypes under two different water treatments. RGB (Red Green Blue) images are processed to estimate projected plant area, which are correlated with destructively measured plant shoot fresh weight (FW), dry weight (DW) and leaf area. Estimated plant FW and DW, along with pot weights, are used to derive daily plant water consumption and water use efficiency (WUE) of the individual plants. Hyperspectral images of plants are processed to extract plant leaf reflectance and correlate with leaf water content (LWC). Strong correlations are found between projected plant area and all three destructively measured plant parameters (R2 \u3e 0.95) at early growth stages. The correlations become weaker at later growth stages due to the large difference in plant structure between the two maize genotypes. Daily water consumption (or evapotranspiration) is largely determined by water treatment, whereas WUE (or biomass accumulation per unit of water used) is clearly determined by genotype, indicating a strong genetic control of WUE. LWC is successfully predicted with the hyperspectral images for both genotypes (R2 = 0.81 and 0.92). Hyperspectral imaging can be a very powerful tool to phenotype biochemical traits of the whole maize plants, complementing RGB for plant morphological trait analysis