Detection of senescence’s dynamics in crop breeding is time consuming and needs considerable details regarding its rate of progression and intensity. Normalized difference red-edge index (NDREI) along with four other spectral vegetative indices (SVIs) derived from unmanned aerial vehicle (UAV) based spatial imagery, were evaluated for rapid and accurate prediction of senescence. For this, 32 selected winter wheat genotypes were planted under full and limited irrigation treatments. Significant variations for all five SVIs: green normalize difference vegetation index (GNDVI), simple ratio (SR), green chlorophyll index (GCI), red-edge chlorophyll index (RECI), and normalized difference red-edge index (NDREI) among genotypes and between treatments, were observed from heading to late grain filling stages. The SVIs showed strong relationship (R² = 0.69 to 0.78) with handheld measurements of chlorophyll and leaf area index (LAI), while negatively correlated (R² = 0.75 to 0.77) with canopy temperature (CT) across the treatments. NDREI as a new SVI showed higher correlations with ground data under both treatments, similarly as exhibited by other four SVIs. There were medium to strong correlations (r = 0.23–0.63) among SVIs, thousand grain weight (TGW) and grain yield (GY) under both treatments. Senescence rate was calculated by decreasing values of SVIs from their peak values at heading stage, while variance for senescence rate among genotypes and between treatments could be explained by SVIs variations. Under limited irrigation, 10% to 15% higher senescence rate was detected as compared with full irrigation. Principle component analysis corroborated the negative association of high senescence rate with TGW and GY. Some genotypes, such as Beijing 0045, Nongda 5181, and Zhongmai 175, were selected with low senescence rate, stable TGW and GY in both full and limited irrigation treatments, nearly in accordance with the actual performance of these cultivars in field. Thus, SVIs derived from UAV appeared as a promising tool for rapid and precise estimation of senescence rate at maturation stages.

Grain yield is increasingly affected by climate factors such as drought and heat. To develop resilient and high-yielding cultivars, high-throughput phenotyping (HTP) techniques are essential for precise decisions in wheat breeding. The ability of unmanned aerial vehicle (UAV)-based multispectral imaging and ensemble learning methods to increase the accuracy of grain yield prediction in practical breeding work is evaluated in this study. For this, 211 winter wheat genotypes were planted under full and limited irrigation treatments, and multispectral data were collected at heading, flowering, early grain filling (EGF), and mid-grain filling (MGF) stages. Twenty multispectral vegetation indices (VIs) were estimated, and VIs with heritability greater than 0.5 were selected to evaluate the models across the growth stages under both irrigation treatments. A framework for ensemble learning was developed by combining multiple base models such as random forest (RF), support vector machine (SVM), Gaussian process (GP), and ridge regression (RR). The R² values between VIs and grain yield for individual base models were ranged from 0.468 to 0.580 and 0.537 to 0.598 for grain yield prediction in full and limited irrigation treatments across growth stages, respectively. The prediction results of ensemble models were ranged from 0.491 to 0.616 and 0.560 to 0.616 under full and limited irrigation treatments respectively, and were higher than that of the corresponding base learners. Moreover, the grain yield prediction results were observed high at mid grain filling stage under both full (R² = 0.625) and limited (R² = 0.628) irrigation treatments through ensemble learning based stacking of four base learners. Further improvements in ensemble learning models can accelerate the use of UAV-based multispectral data for accurate predictions of complex traits like grain yield in wheat.

Environmental stresses from climate change can alter source–sink relations during plant maturation, leading to premature senescence and decreased yields. Elucidating the genetic control of natural variations for senescence in wheat (Triticum aestivum) can be accelerated using recent developments in unmanned aerial vehicle (UAV)-based imaging techniques. Here, we describe the use of UAVs to quantify senescence in wheat using vegetative indices (VIs) derived from multispectral images. We detected senescence with high heritability, as well as its impact on grain yield (GY), in a doubled-haploid population and parent cultivars at various growth time points (TPs) after anthesis in the field. Selecting for slow senescence using a combination of different UAV-based VIs was more effective than using a single ground-based vegetation index. We identified 28 quantitative trait loci (QTL) for vegetative growth, senescence, and GY using a 660K single-nucleotide polymorphism array. Seventeen of these new QTL for VIs from UAV-based multispectral imaging were mapped on chromosomes 2B, 3A, 3D, 5A, 5D, 5B, and 6D; these QTL have not been reported previously using conventional phenotyping methods. This integrated approach allowed us to identify an important, previously unreported, senescence-related locus on chromosome 5D that showed high phenotypic variation (up to 18.1%) for all UAV-based VIs at all TPs during grain filling. This QTL was validated for slow senescence by developing kompetitive allele-specific PCR markers in a natural population. Our results suggest that UAV-based high-throughput phenotyping is advantageous for temporal assessment of the genetics underlying for senescence in wheat.

Cotton is a pivotal global commodity underscored by its economic value and widespread use. In the face of climate change, breeding resilient cultivars for variable environmental conditions becomes increasingly essential. However, the process of phenotyping, crucial to breeding programs, is often viewed as a bottleneck due to the inefficiency of traditional, low-throughput methods. To address this limitation, this study utilizes hyperspectral remote sensing, a promising tool for assessing crucial crop traits across forty cotton varieties. The results from this study demonstrated the effectiveness of four vegetation indices (VIs) in evaluating these varieties for water-use efficiency (WUE). The prediction accuracy for WUE through VIs such as the simple ratio water index (SRWI) and normalized difference water index (NDWI) was higher (up to R² = 0.66), enabling better detection of phenotypic variations (p < 0.05) among the varieties compared to physiological-related traits (from R² = 0.21 to R² = 0.42), with high repeatability and a low RMSE. These VIs also showed high Pearson correlations with WUE (up to r = 0.81) and yield-related traits (up to r = 0.63). We also selected high-performing varieties based on the VIs, WUE, and fiber quality traits. This study demonstrated that the hyperspectral-based proximal sensing approach helps rapidly assess the in-season performance of varieties for imperative traits and aids in precise breeding decisions.