Preharvest phenotypic prediction of grain quality and yield of durum wheat using multispectral imaging

Abstract

Durum wheat is an important cereal that is widely grown in the Mediterranean basin. In addition to high yield, grain quality traits are of high importance for farmers. The strong influence of climatic conditions makes the improvement of grain quality traits, like protein content, vitreousness, and test weight, a challenging task. Evaluation of quality traits post-harvest is time-and labor-intensive and requires expensive equipment, such as near-infrared spectroscopes or hyperspectral imagers. Predicting not only yield but also important quality traits in the field before harvest is of high value for breeders aiming to optimize resource allocation. Implementation of efficient approaches for trait prediction, such as the use of high-resolution spectral data acquired by a multispectral camera mounted on unmanned aerial vehicles (UAVs), needs to be explored. In this study, we have acquired multispectral image data with an 11-band multispectral camera mounted on a UAV and analyzed the data with machine learning (ML) models to predict grain yield and important quality traits in breeding micro-plots. Combining 11-band multispectral data for 34 cultivars and 16 environments allowed to develop ML models with good prediction capability. Applying the trained models to test sets explained a considerable degree of phenotypic variance with good accuracy showing r squared values of 0.84, 0.69, 0.64, and 0.61 and normalized root mean squared errors of 0.17, 0.07, 0.14, and 0.03 for grain yield, protein content, vitreousness, and test weight, respectively.

Crop phenotyping in a context of global change: What to measure and how to do it

“…In this sense, a wide variety of models, both supervised and unsupervised, are applied (Table 3), with the prominent models being partial least squares regression (PLS), least absolute shrinkage and selection operator (LASSO), random forest (RF), support vector machines (SVM), and neural networks (NN). The high potential for application of these methods to HTPP data has been shown in various studies focusing on different aspects of plant science like breeding for plant stress resilience, or yield and quality trait prediction (Naik et al, 2017;Wiegmann et al, 2019;Marques Ramos et al, 2020;Vatter et al, 2022). As all models have benefits and drawbacks it is advisable to evaluate a range of machine learning models for their suitability for the specific experimental trial.…”

Source-Sink Dynamics in Field-Grown Durum Wheat Under Contrasting Nitrogen Supplies: Key Role of Non-Foliar Organs During Grain Filling

“…High-throughput phenotyping has been used to predict yield-related traits in wheat, e.g., using unmanned aerial systems, multispectral cameras, and spectroradiometers (Prey and Schmidhalter, 2020;Garriga et al, 2021;Vatter et al, 2021), but their high cost and user training are drawbacks for their expansion. The vegetative indices used here quantified the greenness by counting pixels in the green-color range or by the spectrum reflected by the vegetation, so their high correlation with biomass and hence GY was not surprising (Figure 4).…”

Advancements, limitations and challenges in hyperspectral imaging for comprehensive assessment of wheat quality: An up-to-date review

“…Test weight is used to reflect the bulk density of wheat grain ( Wang & Fu, 2020 ). Among the four indexes, only the grain yield and hardness were well predicted by HIT (R 2 P > 0.80), while the prediction of vitreousness and test weight by HIT was not satisfactory (R 2 P < 0.65) ( Erkinbaev et al, 2019 , Vatter et al, 2022 ). Detailed information can be found in Table 1 .…”