The Reasearch On The Hyperspectral Monitoring Model Of LAI And Leaf Nitrogen And Yield In Spring Maize

Hyperspectral technology is one of the most important technological means of precision agriculture. It is largely used on plant growth monitoring, plant water monitoring, nutritional status monitoring, yield evaluation and quality monitoring, because of its convenience, high efficiency and no break. Maize is the main part in this study. Analyzing response regulation of maize canopy and leaf hyperspectral features and it’s LAI, SPAD, leaf nitrogen amount and yield under different cultivation conditions (different cultivars, different densities, different amount of nitrogen, density-nitrogen). Clear sensitive band of the four physicochemical indexes above. Utilizing spectral index:NDVI, RVI and DVI to establish LAI, SPAD, leaf nitrogen amount and yield model based on hyperspectral index. The results are as follows:1> The change of the cultivation environment will cause changes of maize physiological parameters, and this changes are discrepant because of the different cultivation measures. Such as:growth period(leaf senescence), cultivation density and amount of nitrogen, etc, can make LAI change, cultivation density effect most, leaf senescence is the second, amount of nitrogen is the last. Similarly, different cultivation conditions also affect canopy and leaf hyperspectral features. On the visible light of 350-760nm, canopy spectra and leaf spectral reflectivity gradually enlarged with growing process went by. Under different cultivation densities, canopy spectral reflectivity decreased with increased density, while leaf spectral reflectivity increased. Under different amounts of nitrogen, canopy spectra and leaf spectral reflectivity increased with high amount of nitrogen. On the visible light of 780-1300nm, canopy spectra and leaf spectral reflectivity increased with growing process. Under different cultivation densities, canopy spectral reflectivity increased with increased density, while leaf spectral reflectivity behaves no definite trend. Under different amounts of nitrogen, canopy spectral reflectivity decreased with high amount of nitrogen, while leaf spectral reflectivity behaves no definite trend. Cultivation density affects canopy spectral reflectivity more than the amount of nitrogen, while the amount of nitrogen affect leaf spectra more than cultivation density.2、By correlation analysis between maize canopy, leaf spectra and LAI, SPAD, LNC, yield under different cultivation conditions, getting sensitive wave length of the four indexes reflectivity are mainly located 550nm,678nm,710nm and 1100nm nearby. First derivative spectra sensitive wave length located 500nm,550nm,580-680nm,700nm and 755nm nearby. Utilizing spectral parameters:NDVI, RVI and DVI to establish LAI, SPAD, LNC and yield model under different cultivation conditions, and comparing model application accuracy:when the model uses other cultivation conditions, there is a big deviation. On canopy spectra model, estimating precision of all the indexes is weak, especially LAI. While the leaf SPAD and LNC model have universality.3、Under different cultivation conditions, discrepancy between hyperspectral parameters and all indexes is the root cause, which causes weak universality in high spectral estimation model. Composite analysis on spectra and all indexes can reduce non-matching level and increase universality. At the same time, utilizing difference between soil parameters leaf spectra and canopy spectral reflectance can make spectral model more stable. On the models above, the better models in universality are:LAI is NDVI model(729.3,963.6), and MNDVI model(729.3,963.3). SPAD and LNC is NDVI model(729.3,963.6), and mRVI model(729.3, 963.6). Yield is NDVI model (R’695.7,R’755.5),and MRVI(550.2,963.6).