Optimization of UAS‐based high‐throughput phenotyping to estimate plant health and grain yield in sorghum

Abstract

High-throughput phenotyping (HTP) has enabled the acquisition of vast amounts of data. Therefore, finding the most informative phenological stage(s) and high-throughput traits could lead to significant optimization of HTP-assisted selection. An investigation as to when phenotypic data should be collected and how it should be processed from unmanned aerial system (UAS) imagery for the optimization and assessment of two primary traits in grain sorghum [Sorghum bicolor (L). Moench], namely, grain yield and plant health (based on anthracnose scores) was conducted. By evaluating multiple flight dates across the growing season via multispectral UAS-based imagery, a set of scenarios composed of combinations of flight dates and vegetation indices were constructed for analysis. In this sense, results showed no increase in predictive ability when combining multiple vegetation indices. Hence, using only an index with a higher predictive ability (e.g., normalized difference vegetation index (NDVI) or modified simple ratio (MSR) for plant health with 0.75; and any tested index but chlorophyll index (CIg) for grain yield with ∼0.55) is recommended. Likewise, the combining of multiple flights did not result in a significant increase in predictive ability for either primary trait. Thus, we observed that a single flight for each trait (e.g., 121 d after sowing with 0.81 for plant health; 104 d after sowing with 0.59 for grain yield) was optimal. Concerning, the predictive algorithms examined, partial least squares regression (PLSR) and neural network, results were similar, with PLSR generally outperforming. In addition, we discuss our findings from an application standpoint of a field-based breeding program and suggest additional optimization options. Abbreviations: AUDPC, area under the disease progress curve; CIg, chlorophyll index; CIre, red-edge chlorophyll index; DAS, days after sowing; GY, grain yield; HTP, high-throughput phenotyping; ML, machine learning; MSR, modified simple ratio; MSRre, red-edge modified simple ratio; NDVI, normalized difference vegetation index; NDVIre, red-edge normalized difference vegetation index; PLSR, partial least squares regression; UAS, unmanned aerial system. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

EnvRtype: a software to interplay enviromics and quantitative genomics in agriculture

“…Combined with phenotyping and genotyping data, the use of envirotyping data may leverage molecular breeding strategies to understand historical trends and cope with future scenarios of environmental change ( Gillberg et al 2019 ; de los Campos et al 2020 ). Its use can also support other prediction-based pipelines in plant breeding, such as high-throughput phenotyping surveys ( Bustos-Korts et al 2019 ; Krause et al 2019 ; Galli et al 2020 ).…”

A low-cost greenhouse-based high-throughput phenotyping platform for genetic studies: a case study in maize under inoculation with plant growth-promoting bacteria

“…The most laborious steps were to manually identify each plot and adjust its shapefile to avoid overlapping plots. Several approaches have been proposed to automate the plot identification step, such as the fieldShape function in the FIELDimageR R package (Matias et al, 2020) or a negative buffer area (Galli et al, 2020). However, these methods did not produce adequate results in our case, probably because of leaf overlapping (Ahmed et al, 2019).…”

Bottlenecks and opportunities in field-based high-throughput phenotyping for heat and drought stress

“…Alternatively, current high-throughput estimations of yield are derived through the analysis of vegetation indices. While this method can provide relatively accurate prediction of yield, it is a secondary measure of yield, and the reliability of the prediction only increases near maturity and could vary based on environmental changes ( Galli et al , 2020 ). In the near future, the primary measurement based on remote sensing that is being developed ( Fig.…”