Evaluation And Modelling Of Portable Crop Nitrogen Monitoring Instrument

Information agriculture which bases on collection technology of information, decision support technology and intelligent equipment to carry out quantitative decision and variable investment for agricultural production with executing in accurate position is the production management technical system of modern agriculture. Furthermore, farmland collection technology of information and equipment are basic and sustain of accurate agriculture and important guarantee of realization high efficiency, high yield and high quality agriculture. Remote sensing technology could be easily used to monitor and forecast growth characters, nitrogen status, yield and quality formation of crop plants, thus it was very important to the development of information agriculture. Non-destructive monitoring technique of crop nitrogen nutrition is not only an urgent need for the key technology for accurate diagnosis and dynamic regulation of crop nitrogen nutrition based on spectral information, but also a technology foundation of nitrogen nutrition non-destructive monitoring equipment development in digital agricultural research.A self-developed portable non-destructive monitoring instrument to measure crop nitrogen nutrition based on crop canopy reflectance spectral is presented. It can measure the incident and reflected radiance of vegetation using the Sun as light source at the 546nm,660nm,710nm,810nm characteristics bands by the 8 specially designed interference filters, then Spectral reflectance values are derived on the LCD converted by single-chip microcomputer. It is portable, simple with less cost and low power assumption.By using self-developed portable instrument for crop nitrogen monitoring, the information of rice canopy reflectance spectra was obtained from the field experiments. The change patterns of canopy spectral reflectance under varied nitrogen rates, different water potential and different growth stages were analyzed. Results showed that the sequence of canopy reflectance spectrum of four characteristic bands included in the instrument was R-810>R546>R-710>R660,and reflectance increased with increasing nitrogen supply. Reflectance at visible band (546.660 and 71 Onm) decreased gradually with the increase in soil moisture content, whereas reflectance in near infrared (810nm) increased. Compared with the same band of Cropscan MSR-16. there was a big difference in spectral reflectance, and variation coefficient of canopy reflectance under different N fertilization levels were smaller integrally.The change patterns of the five parameters (LNC, PNC, LNA, PNA and LAI) with growth periods under varied nitrogen rates were investigated by characterizing the correlations and the quantitative relationships between the agronomic parameters and canopy reflectance in rice. The five parameters were on the rise with the amount of nitrogen increased. The canopy reflectance spectrum of two characteristic bands (660 and 710nm) included in the instrument were negatively correlated with LNC. PNC and LNA, but positively between the five parameters and the other bands (546nmand 810nm).Among the four bands, The correlation coefficients of LNC and PNC with the reflectance at 710nm were-0.78 with a significance testing (p<0.01). By synthesizing the result of the multiple linear regression analysis and curve fitting analysis, we can know that the second order curve coefficients of determination (R2) between the five parameters and the four mono-bands were more than 0.65. The determination coefficients between the five parameters and related bands were 0.78、0.76、0.66、0.50 and 0.55. This study results showed that its ability in monitoring multiple agronomic parameters was excellent, and its performance was steady in field condition.On the base of evaluating the performance of portable nitrogen monitoring instrument. The relationships of leaf dry weight and LAI to ground-based canopy reflectance and spectral parameters were determined. Results showed that leaf dry weight and LAI in wheat could be monitored by the same vegetation indices NDVI(810.710) under varied experiments conditions. Testing of the monitoring models with an independent dataset indicated that the spectral indices gave accurate growth estimation.