Joint prediction of potato yield and nitrogen status at row closure using multi-task deep learning and UAV-derived LiDAR–multispectral data

Key Points

• This research aims to enhance predictions of potato yield and nitrogen status at row closure using UAV-derived data.
• Utilized fused UAV multispectral and LiDAR data to assess potato canopy and nitrogen levels at closure.
• Employed multi-task deep learning models for joint prediction of yield and nitrogen status.
• Conducted feature attribution analyses to explore relationships between canopy structure and nitrogen status.
• Joint modelling improved predictive performance over single-task methods.
• LiDAR metrics provided additional insights beyond multispectral data for nitrogen status.
• Single-date UAV data effectively predicted nitrogen status but were less informative for yield estimation.