Maize Growth Monitoring And N Status Diagnosis Based On Digital Image Technology

The development of portable crop growth monitoring and nutritional diagnosis methods can effectively monitor the growth and nutritional changes of maize during the growth period,achieve precise fertilization,improve field management,and quickly complete the nutrition management of integrated of water and fertilizer in drip irrigation during maize growth period.In order to solve the problems of poor field management accuracy,difficult monitoring of maize growth,and slow dynamic regulation of nitrogen fertilizer in Ningxia Yellow River Diversion Irrigation Area under drip irrigation maize.This study uses the different nitrogen fertilizer gradients N0(0kg·hm-2),N1(70kg·hm-2),N2(140kg·hm-2),N3(210kg·hm-2),N4(280kg·hm-2),N5(350kg·hm-2)digital images of the canopy obtained during each growth period of maize,common color spaces were extracted through segmentation,transformation and other processing,and compared with the measured plant growth and physiological indicators,the characteristic parameter indicators that could invert the growth status and nitrogen nutrition status of maize were screened to verify whether they could stably judge the growth and nutritional status of maize.At the same time,the image feature parameters were used to estimate maize yield.The feasibility of applying digital image technology to the growth,nutrition monitoring and yield prediction under drip-irrigated maize was verified,and a theoretical reference was provided for maize production and nutrition monitoring in Ningxia.The main findings are as follows:(1)Analysis of maize growth monitoring found that maize Hp,LAI,SPAD and characteristic parameters CC and NPCI were in power function relationship,respectively.The model determination coefficient R2 of Hp and CC constructed was 0.438,the model test index RMSE was 87.12,the nRMSE was 38.12%,and the model accuracy is poor and unstable.The determination coefficient R2 of LAI monitoring model based on CC was 0.827,and the model test indexes RMSE and nRMSE were 0.488 and 14.04%,respectively,indicating good stability of the model.The determination coefficient R2 of the SPAD monitoring model constructed based on NPCI was 0.776,and the model test indexes RMSE and nRMSE were 3.07 and 5.56%,respectively.The model had good accuracy and excellent stability.(2)The monitoring of maize biomass found that AGB increased slowly from V4 to V6,rapidly increased and accumulated from V6 to R3,and then the growth rate slowed down.The coefficient of determination R2 of the biomass estimation model of upper maize filed constructed based on MGRVI was 0.792,the model test index RMSE was 2.42,and the nRMSE was 18.89%,and the model stability was good.(3)The research on nitrogen nutrition diagnosis of maize shows that the characteristic parameter EXG based on the RGB color space and the characteristic parameter S based on the HIS color space are extremely significantly correlated with the PNC,which can be described by a power function.The maize nitrogen application diagnostic model PNC based on EXG wasPNC=0.021EXG1.55,the coefficient of determination R2 was 0.759,the model test index RMSE was 2.87,and the nRMSE was 12.76%;the maize nitrogen application diagnostic model PNC based on S was PNC=8.86×108S-3.60,the coefficient of determination R2 was 0.736,the model checking index RMSE was 2.95,and nRMSE was 13.13%.Based on RGB and HIS color space digital image feature parameters,nitrogen nutrition diagnosis in maize growth period can be carried out,and the model has high precision and good stability.(4)The forecast analysis of maize yield shows that the amount of nitrogen fertilization in maize had a significant impact on yield and yield components.Appropriate nitrogen fertilization could increase ear length and ear diameter,reduce bald tip length,and increase 100-kernel weight of maize.Based on maize growth and biomass monitoring model,SPAD*,LAI*,PNC*,AGB*and Y show a quadratic function relationship with one variable.The model R2>0.797 was constructed for each parameter,and the fitting accuracy is high;when the yield reaches the maximum value of maize for V4,V6,V12,R1,R3,R6,the NPCI thresholds were 0.24,0.28,0.30,0.30,0.26,0.23;the CC thresholds were 0.23,0.44,0.64,0.69,0.50,0.42;the EXG thresholds were 121.63,110.76,97.58,88.98,92.40,78.57;MGRVI critical values were 0.17,0.14,0.11,0.10,0.09,0.09.When the monitoring value is lower than the critical value,nitrogen fertilizer should be supplemented in time.Otherwise,nitrogen application is not required.Under N4 treatment,the maximum yield was 13043.01 kg·hm-2,that is,under this research background,the optimal nitrogen application rate was 280 kg·hm-2.