2D Orthomosaic and 3D Modeling Applications in Winter Wheat High-Throughput Phenotyping.

Perhaps the greatest challenge of plant science and crop improvement in the 21st century is predicting how a plant’s appearance (phenotype) is dictated by its genetic make-up (genotype). Advances in “next generation” DNA sequencing are rapidly reducing the costs of genotyping. In contrast, most of the plant phenotypic traits are still extracted using destructive sampling methods and manual measurement, which can be time-consuming, labor-intensive and costly. Recognition of the phenotyping constraint has stimulated development of high-throughput phenotyping (HTP) that seeks to accurately characterize large numbers of individuals or populations using a fraction of time and labor of manual phenotyping methods.  An increasing number of scientists are turning to the use of remote sensing for characterizing plant phenotypes such as plant height.  Generally speaking, remote sensing is associated with ground platforms (e.g. hand-held spectroradiometer) and aerial platforms (e.g. airplane, small unmanned aircraft system or sUAS). For this study, we used a Canon T4i® modified color infrared (CIR) camera carried on a DJI s1000 Octocoper sUAS to build 2D orthomosaics and 3D models of a winter wheat field across the growing season. The objective is to develop the workflow of data processing to extract Normalized Difference Vegetation Index (NDVI) values and canopy height in a reliable, affordable and fast way.

Keywords: sUAS, HTP, canopy height