In a groundbreaking advance for agricultural science and precision breeding, researchers have unveiled a state-of-the-art approach to phenotyping wheat plant height using ultra-low altitude unmanned aerial vehicle (UAV) imagery combined with sophisticated three-dimensional (3D) canopy modeling. This novel methodology leverages low-cost UAV cross-circling oblique (CCO) imaging to generate highly detailed, multi-level volumetric reconstructions of wheat canopies, surpassing traditional nadir-based imaging techniques. By extracting plant height data across multiple quantiles instead of relying solely on average height measurements, the method captures subtle intra-plot variability and yields robust genetic insights. This represents a transformative step forward in high-throughput phenotyping and precision agriculture, with far-reaching implications for accelerating wheat genetic improvement.
Wheat (Triticum aestivum L.) serves as a fundamental staple crop, contributing approximately 20% of global caloric intake. The architecture of the wheat plant, particularly its height, plays an instrumental role in determining yield potential and structural stability. An optimal plant height balances biomass accumulation and photosynthetic capacity against risks of lodging, a phenomenon where excessively tall plants topple under environmental stresses such as wind or rain, leading to substantial yield losses. The “Green Revolution” famously harnessed dwarfing genes to reduce plant height and increase harvest index, revolutionizing global crop productivity. Yet modern breeding programs still face the challenge of precisely tuning plant height to local conditions, environments, and climate variability, necessitating novel methods to quantify this complex trait at scale.
Traditional field-based plant height assessments typically involve manual measurement of a limited number of plants within each plot, a laborious and error-prone approach that fails to fully characterize the spatial heterogeneity within plots. This issue is exacerbated by the time sensitivity and logistical difficulty of such operations, translating into delays or inaccuracies in breeding selection cycles. Recent technological advances have fostered the emergence of high-throughput phenotyping platforms, particularly UAVs outfitted with imaging sensors, enabling rapid, repeated, and non-destructive capture of crop structural traits over large experimental fields. However, classic UAV imaging strategies predominantly utilize nadir (top-down) views, which provide limited canopy perspective, particularly in densely planted or tall crops.
The present study, led by Yuntao Ma and Yonggui Xiao at China Agricultural University and the Chinese Academy of Agricultural Sciences, pioneers the use of cross-circling oblique (CCO) UAV imaging flown at ultra-low altitudes to capture wheat canopies. By flight paths circling plots from oblique angles, the system records comprehensive side and top views, yielding dense 3D point clouds that better resolve the vertical and horizontal complexity of the canopy architecture. Conducted under multi-environmental field trials, this methodology allows direct comparison against traditional nadir imaging, with both approaches flown at identical altitudes and overlap settings to ensure fair benchmarking.
Analytical reconstruction of the CCO-derived point clouds produces precise 3D canopy models from which plant height metrics can be extracted at multiple quantile levels, from lower canopy to uppermost spikes. This multi-quantile approach moves beyond simplistic average height estimations and addresses the intrinsic heterogeneity within and between plots. Of note, results demonstrate that the 90th percentile height quantile exhibits the strongest concordance with ground truth field measurements, while lower quantiles frequently underestimate height by calculating stem rather than spike height. The denser and more accurate canopy coverage afforded by CCO imaging is further validated by its superior correlation coefficients and reduced root mean square errors (RMSE) relative to nadir imaging.
Importantly, the high resolution of CCO 3D reconstructions enables visualization of detailed organ-level features, such as individual spikes within wheat plots, offering phenotyping precision unprecedented in field conditions. Although the method shows some limitations in resolving side views when planting density is exceptionally high, the overall data quality supports robust extraction of phenotypic variation critical for genetic analyses. In this study, recombinant inbred line (RIL) populations evaluated under diverse environments exhibited normal distribution patterns for both field-measured and 3D-derived plant heights, with significant correlations across quantiles and exceptional broad-sense heritability values (ranging from 0.775 to 0.982 depending on environment and quantile).
The study’s power becomes most apparent in its genetic mapping results. A comprehensive quantitative trait locus (QTL) analysis across seven environmental conditions identified 106 loci associated with plant height traits measured by both traditional and 3D methods. Among these, 40 loci were common to both approaches, but crucially, 11 loci were consistently identified only by the multi-level 3D height measurements derived from CCO imaging. The discovery of these stable, previously undetectable loci highlights the enhanced genetic resolution afforded by fine-grained phenotyping. Furthermore, two potentially novel loci, designated QPhzj.caas-3A.2 and QPhzj.caas-7A.1, have been successfully converted into Kompetitive Allele Specific PCR (KASP) molecular markers, validated across natural populations, and shown to associate with significant plant height variation under different irrigation regimes.
Candidate gene analyses anchored to these loci have pinpointed important functional genes such as Rht5, a gibberellin-sensitive dwarfing gene located on chromosome 3B, long implicated in height regulation, and TaGL3-5A on chromosome 5A, known for its influence on grain size and weight. These genetic insights are bolstered by the molecular validation via KASP markers, demonstrating the utility of integrating high-resolution phenomics with genomics for marker-assisted selection (MAS). This integration fosters accelerated breeding gains by enabling early and accurate selection for ideotype traits critical to yield and resilience.
The implications of deploying UAV CCO imaging for multi-level 3D plant height measurement extend beyond wheat. The technique’s scalability, cost-effectiveness, and precision position it as a paradigm-shifting tool for phenotyping diverse crops where canopy architecture and height are agronomically important. As such, this approach aligns seamlessly with emerging trends in digital agriculture and precision phenomics, offering researchers and breeders enhanced capacity to dissect complex traits, monitor crop responses to environmental variables, and optimize genetic improvement pipelines.
This pioneering research not only addresses long-standing technical constraints in field-based phenotyping but also establishes a versatile framework for integrating UAV remote sensing, 3D modeling, and quantitative genetics into routine breeding. As agriculture faces mounting challenges from climate change, resource limitations, and growing food demand, innovations like these are essential for unlocking new genetic potentials and tailoring crops to future environments with unprecedented speed and accuracy.
By providing a refined, multi-dimensional perspective of plant height and its genetic underpinnings, the UAV CCO imaging method represents a transformative advance empowering breeders with actionable data and enabling precision selection strategies. Ultimately, this technology promises to accelerate the development of high-yielding, lodging-resistant wheat cultivars, contributing to global food security and sustainable agricultural intensification.
Subject of Research:
Wheat plant height phenotyping and genetic mapping using UAV-based 3D canopy modeling.
Article Title:
Genetic resolution of multi-level plant height in common wheat using the 3D canopy model from ultra-low altitude unmanned aerial vehicle imagery
News Publication Date:
28 February 2025
References:
DOI: 10.1016/j.plaphe.2025.100017
Keywords:
Agriculture, Technology, Biomedical engineering
