| Grain production is the basis of the sustainable development of the society, and the grain quality is gaining more and more attention among increasing yield. Real-time monitor of wheat grain quality can reasonably guide grain classification and achieve quality grain production. Nitrogen is an important element for crops, reasonable application of nitrogen fertilizer is an important means of agriculture production. Unreasonable use of nitrogen fertilizer will reduce nitrogen use efficiency and pollute environment. So monitoring real-time nitrogen deficiency condition is important for scientific fertilization management to increase the yield, improve quality of wheat and protect environment.The relationship between winter wheat agronomic parameters, nitrogen deficit and canopy spectral reflectance of different growing periods under different nitrogen management modes were analyzed to extract spectral sensitive bands, and build agronomy parameter spectrum model and nitrogen deficiency model with the method of multivariate statistical analysis. Moreover, the regression models of grain protein yield of winter wheat were established based on the technical route of "canopy spectral-agronomy parameters-grain protein yield". The conclusions as follow:1〠The canopy spectral curves of winter wheat were similar in the whole growth period under different nitrogen management modes. The reflectance in visible band was reduced and increased in near infrared band with increasing nitrogen level, and red edge parameters and absorption parameters were increased. The difference between the treatment of 300kg · hm-2 and 225 kg ·hm-2 was not obvious. Under the same nitrogen level, the reflectance in visible band of 5:5 (ratio of base and topdressing) was lower and higher in near infrared band than 7:3.2〠LAI of winter wheat was first increased and then decreased, and chlorophyll content was similar to LAI. Dry biomass above ground was increased with the growing time, and plant nitrogen accumulation was similar to it At different groeth neriod of winter wheat they were increased hased on the increased application of nitrogen and their values under the ratio of base and topdressing of 5:5 were higher than 7:3 when the nitrogen application was the same level. The extraction of sensitive bands for LAI, content of chlorophyll and nitrogen accumulation based on PLS-SMLR method were 775,765,1060 nm,675,935, 725,560,865 nm and 550,445,400,725,975,530,840 nm, respectively.3〠The difference of grain protein yield between 300kg · hm-2 and 225 kg · hm-2 was not significant, while both of them were significant compared with the treatment of 150 kg · hm-2 and without nitrogen application. There was same tendency under different ratio of base and topdressing. There were high linear correlation between grain protein yield and LAI, PNA in filling stage, and the linear correlation between grain protein yield and the content of chlorophyll in booting stage was highest. The spectral estimation model of grain protein yield based on agronomy parameter of PNA was established with the highest R2 and lowest RMSE, RE. It indicated that grain protein yield of winter wheat could be predicted with its canopy spectral reflectance at filling stage and the result would guide crop production.4〠Nitrogen nutrition index(NNI) was increased and the accumulation nitrogen deficit (Nand) was decreased gradually with the increased application of nitrogen in the main growth period before booting. The NNI under the same nitrogen level 5:5 ration of base and topdressing treatment was higher than 7:3, while there was a reverse tendency for Nand at jointing, heading and booting stage of winter wheat. It was 440nm and 600nm that were the spectral characteristic bands to monitor Nand of the winter wheat. Monitoring model was established based on the characteristic bands with high R2, low RMSE, RE using PLS-SMLR method. It could guide nitrogen application according to canopy spectral on different growth period of winter wheat, in order to meet the nitrogen needs of crop growth. |