Uav-based high-throughput phenotyping platform for crop breeding programs
Field-ready UAV imaging system combined with machine learning and deep learning analytics that accelerates crop trait measurement, improves selection accuracy, and shortens breeding cycles. Validated across 12,000–25,000 soybean breeding lines and multi-location yield trials over four years.
Overview
This solution offers a field-based high-throughput phenotyping platform that integrates unmanned aerial vehicle (UAV) imaging, machine learning, and deep learning analytics to strengthen conventional crop breeding programs. By capturing spectral and thermal imagery across large breeding populations, the system enables faster measurement of key agronomic traits, improves selection accuracy, and shortens the breeding cycle. The platform has been actively developed and validated in soybean breeding programs, supporting yield prediction, stress response quantification, and automated field note-taking for traits such as maturity date, plant height, flowering time, and emergence.
Technical specifications
Key features:
- UAV-based imaging payloads combining spectral and thermal sensors for rapid field-level data collection
- Machine learning and deep learning pipelines that convert raw imagery into quantitative crop traits and yield predictions
- Scalable processing workflows tested on breeding populations of 12,000 to 25,000 lines, including F5 progeny, multi-location yield trials, and production fields
- Automated trait extraction covering maturity, plant height, flowering time, emergence, and responses to abiotic and biotic stress
- Published methodologies validated through peer-reviewed research in collaboration with breeders
Collaboration needs:
- High-performance computing resources for large-scale image and data analytics
- Software development support to harden academic pipelines into deployable tools
- Connections to breeding programs for side-by-side field validation
- Funding for key personnel to advance system development
Technology readiness level
The phenotyping system and analytics pipelines have been tested and refined over four years within collaborative soybean breeding programs, demonstrating reliable performance across large populations and multiple field locations. To transition from an academic prototype to a production-ready tool, the team is conducting side-by-side comparisons with breeders to benchmark efficiency and effectiveness, while continuing to improve deep learning algorithms and overall system performance. The platform is positioned at an advanced validation stage, ready for pilot deployment and co-development with breeding organizations and technology partners.
About University of Missouri, Columbia
Founded as Missouri’s flagship land‑grant, the University of Missouri–Columbia is a comprehensive public research university serving a large student body and partners statewide. Industry engages on campus through co‑located research cores, a research park and incubator in Columbia, and access to a high‑power university research reactor supporting isotope production and advanced testing. An integrated academic health system enables clinical studies and translation, while a statewide extension and agricultural research network links companies to field sites, producers, and community testbeds across Missouri. Research is supported by competitive federal funding from agencies such as NIH, NSF, USDA, and DOE. A dedicated tech transfer office supports IP, licensing, and sponsored‑research agreements.
delivered to your inbox.