High-throughput phenotyping has become the frontier to accelerate breeding through linking genetics to crop growth estimation, which requires accurate estimation of leaf area index (LAI). This study developed a hybrid method to train the random forest regression (RFR) models with synthetic datasets generated by a radiative transfer model to estimate LAI from UAV-based multispectral images. The RFR models were evaluated on both (i) subsets from the synthetic datasets and (ii) observed data from two field experiments (i.e., Exp16, Exp19). Given the parameter ranges and soil reflectance are well calibrated in synthetic training data, RFR models can accurately predict LAI from canopy reflectance captured in field conditions, with systematic overestimation for LAI<2 due to background effect, which can be addressed by applying background correction on original reflectance map based on vegetation-background classification. Overall, RFR models achieved accurate LAI prediction from background-corrected reflectance for Exp16 (correlation coefficient () of 0.95, determination coefficient () of 0.90~0.91, root mean squared error (RMSE) of 0.36~0.40 m² m^-2, relative root mean squared error (RRMSE) of 25~28%) and less accurate for Exp19 ( =0.80~0.83, = 0.63~0.69, RMSE of 0.84~0.86 m² m^-2, RRMSE of 30~31%). Additionally, RFR models correctly captured spatiotemporal variation of observed LAI as well as identified variations for different growing stages and treatments in terms of genotypes and management practices (i.e., planting density, irrigation, and fertilization) for two experiments. The developed hybrid method allows rapid, accurate, nondestructive phenotyping of the dynamics of LAI during vegetative growth to facilitate assessments of growth rate including in breeding program assessments.

Ground coverage (GC) allows monitoring of crop growth and development and is normally estimated as the ratio of vegetation to the total pixels from nadir images captured by visible-spectrum (RGB) cameras. The accuracy of estimated GC can be significantly impacted by the effect of ‘mixed pixels’, which is related to the spatial resolution of the imagery as determined by flight altitude, camera resolution and crop characteristics (fine vs coarse textures). In this study, a two-step machine learning method was developed to improve the accuracy of GC of wheat (Triticum aestivum L.) estimated from coarse-resolution RGB images captured by an unmanned aerial vehicle (UAV) at higher altitudes. The classification tree-based per-pixel segmentation (PPS) method was first used to segment fine-resolution reference images into vegetation and background pixels. The reference and their segmented images were degraded to the target coarse spatial resolution. These degraded images were then used to generate a training dataset for a regression tree-based model to establish the sub-pixel classification (SPC) method. The newly proposed method (i.e. PPS-SPC) was evaluated with six synthetic and four real UAV image sets (SISs and RISs, respectively) with different spatial resolutions. Overall, the results demonstrated that the PPS-SPC method obtained higher accuracy of GC in both SISs and RISs comparing to PPS method, with root mean squared errors (RMSE) of less than 6% and relative RMSE (RRMSE) of less than 11% for SISs, and RMSE of less than 5% and RRMSE of less than 35% for RISs. The proposed PPS-SPC method can be potentially applied in plant breeding and precision agriculture to balance accuracy requirement and UAV flight height in the limited battery life and operation time.

To maximize the grain yield of spring wheat, flowering needs to coincide with the optimal flowering period (OFP) by minimizing frost and heat stress on reproductive development. This global study conducted a comprehensive modelling analysis of genotype, environment, and management to identify the OFPs for sites in irrigated mega-environments of spring wheat where the crop matures during a period of increasing temperatures. We used a gene-based phenology model to conduct long-term simulation analysis with parameterized genotypes to identify OFPs and optimal sowing dates for sites in irrigated mega-environments, considering the impacts of frost and heat stress on yield. The validation results showed that the gene-based model accurately predicted wheat heading dates across global wheat environments. The long-term simulations indicated that frost and heat stress significantly advanced or delayed OFPs and shrank the durations of OFPs in irrigated mega-environments when compared with OFPs where the model excluded frost and heat stress impacts. The simulation results (incorporating frost and heat penalties on yield) also showed that earlier flowering generally resulted in higher yields, and early sowing dates and/or early flowering genotypes were suggested to achieve early flowering. These results provided an interpretation of the regulation of wheat flowering to the OFP by the selection of sowing date and cultivar to achieve higher yields in irrigated mega-environments.

Aboveground dry weight (AGDW) and leaf area index (LAI) are indicators of crop growth status and grain yield as affected by interactions of genotype, environment, and management. Unmanned aerial vehicle (UAV) based remote sensing provides cost-effective and non-destructive methods for the high-throughput phenotyping of crop traits (e.g., AGDW and LAI) through the integration of UAV-derived vegetation indexes (VIs) with statistical models. However, the effects of different modelling strategies that use different dataset compositions of explanatory variables (i.e., combinations of sources and temporal combinations of the VI datasets) on estimates of AGDW and LAI have rarely been evaluated. In this study, we evaluated the effects of three sources of VIs (visible, spectral, and combined) and three types of temporal combinations of the VI datasets (mono-, multi-, and full-temporal) on estimates of AGDW and LAI. The VIs were derived from visible (RGB) and multi-spectral imageries, which were acquired by a UAV-based platform over a wheat trial at five sampling dates before flowering. Partial least squares regression models were built with different modelling strategies to estimate AGDW and LAI at each prediction date. The results showed that models built with the three sources of mono-temporal VIs obtained similar performances for estimating AGDW (RRMSE = 11.86% to 15.80% for visible, 10.25% to 16.70% for spectral, and 10.25% to 16.70% for combined VIs) and LAI (RRMSE = 13.30% to 22.56% for visible, 12.04% to 22.85% for spectral, and 13.45% to 22.85% for combined VIs) across prediction dates. Mono-temporal models built with visible VIs outperformed the other two sources of VIs in general. Models built with mono-temporal VIs generally obtained better estimates than models with multi- and full-temporal VIs. The results suggested that the use of UAV-derived visible VIs can be an alternative to multi-spectral VIs for high-throughput and in-season estimates of AGDW and LAI. The combination of modelling strategies that used mono-temporal datasets and a self-calibration method demonstrated the potential for in-season estimates of AGDW and LAI (RRMSE normally less than 15%) in breeding or agronomy trials.