Diagnosis On Nitrogen Status Of Maize Using Spectral Parameter Of Nitrogen Nutrition Indicate Leaf

Nitrogen is one of the important nutrients. It is the necessary matter for yield and quality characteristics of maize due to its involvement in synthesis and transport of photosynthates to plant reproductive and storage organ, and subsequent conversion into amino acid, chlorophyll, enzyme, proteins, and other products. If maize lacked nitrogen nutrition, it can not grow adequately. Otherwise, fertilizer would be wasted and yield and quality would be affected. So it is very important to develop precision variable rate application and manage nitrogen fertilizer effectively. Nitrogen management depends on evaluating nitrogen status of soil and crop accurately, therefore real-time、fast and accurate evaluation of maize nitrogen status has a scientific and practical value to rational fertilizer application In this study, maize was selected as research object under different nitrogen levels at different growth stages. Difference of leaf spectral reflectance under different frequency among treatments and correlation relationship between leaf nitrogen content of different leaf position and plant nitrogen content were analyzed and nitrogen nutrition indicates leaf sensitive was obtained. Spectral parameters of nitrogen nutrition indicates leaf models for maize plant nitrogen nutrition diagnosis were constructed. At last, combined with plant testing for maize nitrogen status diagnosis and fertilizer recommendation, the possibility of using the spectral parameter of the maize nitrogen nutrition indicates leaf to detect maize nitrogen status and make N fertilizer recommendation studied.Its primary conclusion were as follows:1、Every leaf nitrogen content increased with the increase of nitrogen application levels and the influence of low level leaf position is lower than high level leaf position. Increased nitrogen fertilizer application can increase the difference of leaf nitrogen content between upper and lower leaves. The differences between the nitrogen nutrition indicates leaf and non nitrogen nutrition indicates leaf of spectral parameters R550had the biggest correlation coefficient, which was bigger than the correlation coefficient of R550of the nitrogen nutrition indicates leaf and plant nitrogen content. The difference of spectral parameters between the nitrogen nutrition indicates leaf and other leaf can do maize plant nitrogen nutrition diagnosis well.Leaf nitrogen content of different leaf position was larger differences at jointing stage, booting stage, tasseling stage of maize. Thus to determine the best leaf represents the plant nitrogen nutrition, is the precondition of maize nitrogen nutrition precise diagnosis. The test results show that, the correlation between nitrogen content of6th leaf of jointing stage,12th leaf of tasseling stage and booting stage and plant nitrogen content was best, so three leaf positions of different growth period above can be as nitrogen nutrition indicates leaf respectively. Nitrogen nutrition indicates leaf of spectral parameters were used to establish the linear prediction model can be well estimated on plant nitrogen content in each growth period.2、Different fertilizer application levels influence on the each growth stage of soil nitrate nitrogen content, alkali solution nitrogen content and total nitrogen content obviously, so the rational use of nitrogen fertilizer can improve the ability for nitrogen of soil.The correlation coefficient between different growth stage of soil nitrate nitrogen content, alkali solution nitrogen content and maize plant nitrogen content was between0.73～0.82, and corn plant nitrogen content and total nitrogen content in soil is relatively poor, the every growth stage of the correlation coefficient are below0.5, so soil total nitrogen content as a evaluation index of soil nitrogen supply ability remains to be further consideration.The correlation between soil nitrate nitrogen content, alkali solution nitrogen content and spectral parameters (R550, NIR/G, mND705) of each growth stage of indicator leaf position were better and had reached the significant level, With relatively poor correlation spectrum parameter of SR, PRI. correlation coefficients between soil total nitrogen content and the five spectrum parameters did not reach significant level.3, The amounts of nitrogen fertilizer recommended for top-dressing of different treatments with varied basal N rates were calculated with the newly developed N top-dressing model based on plant nitrogen content, which was estimated from near infrared spectral parameter. Fertilizer N rate followed a diminishing corn yield pattern, the results showed that the one factor and two factor model can simulate the relation well.Thereafter, the critical values of maize plant nitrogen content determined for N fertilizer dressing of jointing stage.4. In order to increase the detection accuracy of corn plant nitrogen nutrition diagnosis, information fusion technology of digital image color feature and spectral parameters was introduced for plant nitrogen nutrition diagnosis. Firstly, the digital image color feature were captured and analyzed, then six variables including the R、G、B、r、g、b were extracted. With the six variables as inputs, BP neural network was built to diagnose plant nitrogen content. Secondly, the spectral parameters were captured and processed, with the five variables including R550、NIR/G、mND705、R、PRI as inputs, BP neural network was developed to diagnose plant nitrogen content. Finally, plant nitrogen content were diagnose based on color feature and spectral parameters. The results showed that the RMSE of plant nitrogen content were2.33～3.11and1.84～3.37respectively by color feature and spectral parameters. However, the RMSE of plant nitrogen content were1.34～2.12by the information infusion of two techniques. The information fusion technology was better than single technique.5、More researches were carried out to investigate the application of spectral technology on crop nutrition diagnosis. However, the complex conditions in the field results in the uncertainty of spectrum. In this paper, the influence of foliar dust on spectral of the crop beside the nation road was studied, the differences of the raw spectral reflectance and first derivative spectral reflectance between the foliar with dust and the foliar washed with deionised water were analyzed, nitrogen prediction models were built on the disturbance of foliar dust. Result showed that the dust foliar spectral reflectance increased in the visible light (350～720nm) and shortwave infrared (SWIR)(1360～2500nm) regions whereas the spectral reflectance in the near infrared (NIR)(720～1360nm) wavelength regions decreased. There were no change rules for blue edge position, yellow edge position, red edge position, blue edge slope and yellow edge slope on the effect of foliar dust, but red edge slope, blue edge area, yellow edge area, red edge area decreased. Foliar dust also does have effect on the model precision. Foliar dust as a condition of spectral detecting field crop should been assessed.