ABSTRACT
Cereal crops (e.g., rice, wheat, and maize) consume more than 50% of the nitrogen (N) produced industrially in the world. The optimal management of N fertilizer in cereal crops creates great demand for the nondestructive and accurate monitoring of crop N status. Hyperspectral remote sensing has been successfully used to estimate the leaf N concentration (LNC) of cereal crops for more than a decade. In this chapter, we will start a review of hyperspectral remote sensing of LNC in cereal crops by describing the physical principles of the estimation of LNC from reflectance data at the leaf and canopy levels. We will then summarize the current status of LNC estimation from existing studies with respect to the crop types of interest, coverage of growth stages, level of observation, and analytical methods, including the commonly used vegetation indices and emerging continuous wavelet-based spectroscopic analysis. This summary will be followed by a detailed discussion of the technical and theoretical challenges that need to be addressed with more effort in terms of the effect of crop growth stage on predictive model consistency, the lack of physical inversion approaches for LNC estimation, and the correction for the confounding effect of canopy structure. Last, we will point out the further considerations of the effective use of reflectance information in the shortwave-infrared region and the synergy of multiangle observations for improved estimation of LNC and describe the outlook for the spatial mapping of LNC over major areas of cereal crop production with the upcoming spaceborne hyperspectral instruments.
