Monitoring Of Plant Type And Grain Protein Quality Based On Hyperspectra In Wheat

Based on the Imminence for obtaining plant type and grain quality information timely in wheat, we selected some cultivars with different plant types. It was analyzed about the canopy spectral characteristic in different plant types and the correlation among chlorophyll content, grain quality indices and spectral characteristic parameters. Furthermore, the different methods detecting wheat nitrogen status were compared using hyperspectral meter. The main results are summarized as below:Canopy spectral reflectance of those wheat cultivars with lax leaves is higher than that with erect leaves in near infrared bands (700-1300nm) at earlier stages (especially at jointing stage), and the difference decreased with the population expanding. It also indicted that normal difference vegetable index (NDVI) of those cultivars with lax leaves is higher than that with erect leaves when vegetable coverage (COV) are equal. Besides, NDVI increased With COV increasing, but the difference between two plant types became diminishing, which reflected their plant-type characteristics. Different plant types can be classified using COV and NDVI based on the significant correlation between NDVI and leaf area index (LAI) before the field was not covered completely, and jointing stage is the best stage to identify plant type.Leaf Orientation Value (LOV) and LAI are two important indices of canopy structure, and the correlation between them and canopy spectral parameters was analyzed in this study. The results indicted that LOV affected greatly canopy spectra more than LAI in jointing stage, but LAI had more effects to spectra after the ground was near to be covered completely. The result of cluster analysis showed that all treatments can be parted into four clusters: A-erect type and low LAI, B-erect type and high LAI, C-horizontal type and low LAI, D-horizontal type and high LAI. Their spectral reflectance had obvious difference in 400-1150nm at jointing, and it showed A>B>C>D in 400-700nm but contrary order in 700-1150nm. It indicted that the combination of plant-type and the population magnitude can be identified qualitatively using the reflectance in 89Qnm band (R890) and R890 dispersion from jointing stage to booting stage (AR890), and the disciminant functions were also established.The leaf chlorophyll content in different leaf positions shows descending from up to down, namely, up-layer (L1) > mid-layer (L2) > sub-layer (L3), but their dispersions were different in different growing stages and plant types. Correlation analysis indicated that the sensitive stage for detecting is in grain filling stage, and the sensitive parameters are dispersion of chlorophyll content between L1 and L2 (DCC), grain protein content (GPC) and P_Depth560 respectively. Based on the correlation among DCC-GPC- P_Depth560, GPC can be predicted using P_Depth560 at gram filling stage, which also provided a feasible way for predicting grain quality indices by remote sensing.Different methods involving detector vertical , horizontal and declining with 10-20° were used to measure wheat variety spectrum in ripe stage by ASD2500. The results show that the spectral reflectance achieved with different methods all had the typical vegetation spectra characteristics and the same change trend in 350-2500nm, yet their spectral reflectance value had obvious difference. Ear-layer spectral reflectance (Rel) was much closer to ear spectral reflectance (Re) than canopy spectral...