The convergence of smart sensing technology, robotics, and artificial intelligence is ushering a new era of digital agriculture to realize the potential of large agricultural datasets to enhance plant breeding, crop management, and sustainable agriculture. Undertaking the big data-driven genotyping-to-phenotyping challenges, high-throughput plant phenotyping (HTPP) can generate high geospatial resolution measurements of plants and enable high temporal resolution measurements through multiple crop growth stages, allowing a dynamic view of the plant phenotypic variation. In the Poland Lab at Kansas State University, we have been testing and implementing optimal ground- and aerial-based HTPP technologies to improve genomic prediction models for yield and agronomic traits prediction. In this presentation, two research projects will be highlighted. The first project provides an example of how to quantify a complex plant trait (i.e., the heading date in wheat) by HTPP and deep learning on “breeder-trained” datasets. The second project demonstrates data analytics for HTPP with unmanned aerial systems. The latter project will be presented along with case studies of sorghum height measurement and wheat lodging assessment. As more tools are available for HTPP to generate larger datasets, implementation of robust and efficient data analysis methods are in high demand to further advance genomic prediction.

Presenter(s)

Xu Wang