description
- This project lays the foundations for next generation remote sensing of crops to determine nitrogen use efficiency (NUE) and associated canopy photosynthetic potential. We will develop a high-throughput novel hyperspectral (HS) analysis technique that directly extracts biological meaningful data from leaves and canopies at different scales, from leaf clip-on sensors, cameras mounted on field robots and/or drones and satellite earth observation (EO). We exploit recent research at ANU which showed HS imaging combined with machine learning can accurately predict NUE, leaf photosynthesis carboxylation rate (Vcmax) and electron transport, and respiratory carbon loss. This information can be directly exploited to accelerate the phenotyping and breeding of crops as well for agronomy and nitrogen application rate support for farmers. This feasibility study will test the approach in both Australia and the UK and considers how and whether hyperspectral imaging from space could add value as a novel EO product for agriculture. Brief overview of the project This project lays the foundations for the use of hyperspectral imaging and remote sensing to estimate crop nitrogen use efficiency and photosynthetic capacity of wheat. The project is grounded on breakthrough discovery science at ANU which used hyperspectral imaging and machine learning to recover the photosynthetic potential of leaves. Here we extend the state of the art to examine whether hyperspectral imaging at a range of scales from sensors attached to leaves, field robots, UAVs or earth observation satellites can retrieve similar biologically relevant information for whole plants and crops. The project enables improved farmer decision support to optimise N use, the phenotyping of crops by plant breeders for high yield traits and derisks future investments in the use of hyperspectral imaging for earth observation. It mitigates the environmental impact of farming, since input nitrogen accounts for highly significant proportions of arable system GHG emissions. The research is a collaboration between the University of Lincoln, STFC-RAL, ANU and University of Sydney. Datasets will be collected in the UK at UoL, NIAB and Rothamsted, complemented by work in the counter cyclical season in ANU. b. Key objectives and expected outcomes To collect high quality hyperspectral imaging and ground truth leaf photosynthetic function and NUE datasets at a range of scales (leaves, robots, UAV's, EO) in both the UK and Australia. To assess whether hyperspectral imaging at a range of sensing scales can recover biologically relevant data on leaf photosynthetic capacity and nitrogen use efficiency. To consider how the information can be applied by farmers, plant breeders and developers of next generation earth observation products. The outcomes are robust and scientifically underpinned assessment of the potential use of hyperspectral imaging in remote sensing as well as key recommendations for its onward use across multiple remote sensing platforms.