DSFAS-AI: HARNESSING MACHINE LEARNING AND HYPERSPECTRAL IMAGING FOR HIGH-THROUGHPUT MAIZE SILAGE PHENOTYPING

The overall goal of this project is to assess maize silage yield and quality traits in a high-throughput manner by melding cutting-edge aerial hyperspectral imaging and machine learning technologies in a field setting.The maize silage mix is more complex than other forage species as it includes grain and stover, therefore methods developed for this plant structure can potentially be applied to other forage commodities with less complex plant structures for biochemical composition assessment.To perform high-throughput maize silage phenotyping and tackle the associated data science challenges, in this project, we plan to pursue the following three specific objectives:1.Assess spectral variability across maize silage breeding populations evaluated across multiple field locations as part of an ongoing breeding program.2.Develop time-series feature extraction and a multi-task machine learning model to estimate silage yield and multiple quality traits and identify the best time of applying the UAV-hyperspectral survey.3.Develop an unsupervised domain adaptation approach to enhance the model generalizability across different environments.