Monitoring Nitrogen Status And Growth Characters With Ground-Air Hyperspectal Remote Sensing In Wheat

Effective monitoring of plant nitrogen status is the foundation for precision management and dynamic regulation in crop fertilization. Remote sensing can timely and nondestructively determine the nitrogen and growth status of crop in the field, which offers important technical support for implementation of precision farming and realizing high yielding, good quality and high efficiency in modern agricultural production. The newly-emerged hyperspectral remote sensing, which has high resolution and consecutive spectra, is sensitive to specific crop parameters, with better estimation of various growth variables related to crop physiology and biochemistry. The run of Hyperion, the first satellite held hyperspectral imaging spectroradiometer, made application of hyperspectral information at large scale become possible.In this study, a series of field experiments with different wheat varieties and nitrogen levels were carried out, multi-plat wheat spectral reflectance were obtained with satellite and ground hyperspectral spectroradiometer, field sampling and testing synchronously implemented. Based on technology of remote sensing image processing and analysis of spectral reflectance and mathematical statistic, the characteristic spectral parameters related to wheat nitrogen status were screened out, the models that monitoring wheat nitrogen status and growth characters based on reflectance spectra were established in this paper. Then, based on the spectral monitoring of plant nitrogen status and the quantitative relationships between phasic leaf nitrogen status and grain protein content, the indirect approach predicting mature grain content with reflectance spectra were developed. The key technology for real-time monitoring of wheat nitrogen and growth status with ground and satellite hyperspectral sensors was established in this study.Crop leaf chlorophyll status is the important index estimating plant photosynthetic efficiency and nutritional stress, timely and non-destructive monitoring of leaf and canopy chlorophyll status are foundations of nitrogen monitoring and growth diagnosis of wheat. Based on the field experiments under different fertilization levels, wheat leaf and canopy spectral reflectance were obtained with field hyperspectral spectroradiometer (over 350-2500nm), quantitative relationships of leaf chlorophyll concentrations to their hyperspectral parameters were analyzed, and monitoring models were established for leaf chlorophyll concentrations in wheat. Characterized geometric patterns of the first derivative reflectance spectra in red edge area under different chlorophyll levels with wheat, a new red edge parameter. defined as red edge symmetry (RES), was developed to monitor leaf chlorophyll status. Compared to different spectral parameters, the RES was more closely and stably related to chlorophyll concentrations. Furthermore the RES could be easily calculated with only two red edge boundary reflectance (R675, R755) and the red edge center wavelength reflectance (R718), with the equation RES=(R718-R675)/(R755-R675).This made the RES readily applicable to the common airborne and satellite hyperspectral data such as AVIRIS and Hyperion sources, besides ground-based spectral reflectance.Crop nitrogen status is the important evaluation index of crop growth status and soil nitrogen level, timely and non-destructive monitoring of leaf nitrogen status is significant to optimize nitrogen management and improve grain yield and quality. This research made a quantitative analysis on the characteristics of the first derivative reflectance spectra in red edge area under different nitrogen levels based on canopy hyperspectral reflectance with field-grown wheat, a new algorithm for fitting REP, named as simplified linear extrapolation REP was constructed. Compared to some common vegetation indices, the REPs derived from simplified linear extrapolation were more closely related to leaf nitrogen concentrations in wheat grown under varied nitrogen rates. The relationship between the REP(SLE) based on simulated wavebands of hyperspectral satellite sensor Hyperion and leaf nitrogen concentrations of wheat showed that the new REP algorithm was readily applicable to the broadband data as well as narrowband data. Case study on real-time Hyperion image further demonstrated that this new formulation could well monitor leaf nitrogen concentrations in the experiment area.Leaf area index (LAI) is the important index of crop growth status, timely and non-destructive monitoring of LAI is significant to growth diagnosis and dynamic regulation in wheat production. Results indicated that wheat LAI increased with the increase of nitrogen fertilizer application rates. A quantitative analysis on the relationship between reflectance spectra and canopy hyperspectral reflectance indicated that near-infrared reflectance was sensitive to wheat LAI. The estimation models based on hyperspectral parameters, FD750, R762.6-R742.25, DVI(810,680) and DVI(810,560) were effective and reliable. Compared to broadband vegetation indices, such as GVI(AVHRR), TC3, the narrowband parameters had not significant advantage in monitoring wheat LAI. Case study on real-time Hyperion image demonstrated that hyperspectral data could well monitor wheat LAI in the experiment area.Based on the field experiments under different fertilization levels, results showed that leaf dry weight increased with the increase of nitrogen fertilizer rates in wheat. Correlations of leaf dry weight and above-ground dry weight to canopy hyperspectral reflectance showed that the former was good while the latter was poor, with the sensitive spectral wavebands located at NIR range. The models based on hyperspectral parameter RVI(810,680) gave high estimation efficiency. Case study on real-time Hyperion image demonstrated that hyperspectral data could well monitor wheat leaf dry weight in the experiment area, with the result dependent on the precision of radiation calibration. Broadband vegetation indices, such as SARVI(MSS), GEMI(AVHRR) were also closely related to wheat leaf dry weight, indicating that it is feasible to monitor wheat leaf dry weight with multi-spectral information.Relationships of leaf nitrogen concentration to grain protein content under different wheat cultivars and growth stages showed that grain protein character at maturity could be forecast by plant nitrogen status at pre-maturity, with anthesis as a proper stage, although this relationship slightly varied with different variety types. Based on the technical route of key spectral parameters-leaf N concentrations-grain protein contents, total predicting models on grain protein content were constructed by linking monitoring model on leaf nitrogen status with hyperspectral remote sensing and predicting model on grain protein content based on leaf nitrogen concentration. Testing of the predicting models indicated that the spectral indices of REP(SLE) at anthesis gave accurate estimation of grain protein contents in wheat. Then, the grain protein contents were further predicted using real-time Hyperion image in the wheat experiment area.