Associate Professor North Carolina State University
Disclosure: Disclosure information not submitted.
Spectral reflectance imaging has been used extensively as a rapid means of assessing certain aspects of a plant’s metabolome. However, as high throughput phenotyping moves to higher spatial and temporal resolutions, confounding effects due to a leaf’s bidirectional reflectance are becoming more glaring. This effect, which is quantified using a bidirectional reflectance distribution function (BRDF), causes significant view- and illumination-angle dependencies within detected spectra. While BRDF correction and modeling methods exist (e.g., SAIL/PROSAIL), they can be complex to invert, requiring a significant investment of both time, effort, and computing power, or they require greater amounts of data (e.g., both space and spectrum combined with time or angle). In this presentation, we will overview our progress towards developing an emerging method of performing BRDF corrections using imaging polarimetry. This will include an overview of full-season diurnal field trial measurements, taken using our hyperspectral imaging polarimeter, as well as laboratory polarization-BRDF (pBRDF) measurements of maize leaves taken using our hyperspectral Mueller matrix pBRDF platform. Through this work, we demonstrate the potential of imaging polarimetry to play a supporting role in correcting BRDF-related effects in hyperspectral imaging, as deployed within high throughput phenotyping scenarios.