基于作物生长监测诊断仪的双季稻叶片氮含量和氮积累量监测

为了验证作物生长监测诊断仪(CGMD)监测双季稻氮素营养指标的准确性和适用性,构建基于CGMD的双季稻叶片氮含量(LNC)和氮积累量(LNA)的监测模型。选用8个不同早、晚稻品种,设置4个不同施氮水平,利用CGMD采集冠层差值植被指数(DVI)、归一化植被指数(NDVI)和比值植被指数(RVI),同步利用ASD FH2高光谱仪采集冠层光谱反射率,并计算DVI、NDVI和RVI;通过比较CGMD和ASD FH2采集的冠层植被指数变化特征,验证CGMD的测量精度,构建基于CGMD的LNC和LNA监测模型,并利用独立试验数据对模型进行检验。结果表明: 早、晚稻LNC、LNA、DVI、NDVI和RVI随施氮水平的增加而增大,随生育进程的推进呈先升后降的趋势;CGMD与ASD FH2采集的DVI、NDVI和RVI间拟合的决定系数(R²)分别为0.9350、0.9436和0.9433,表明CGMD的测量精度较高,可替代ASD FH2采集冠层植被指数。基于CGMD的3个冠层植被指数相比,NDVI_CGMD与LNC的相关性最高,RVI_CGMD与LNA的相关性最高;基于NDVI_CGMD的指数模型可较准确地预测LNC,模型R²为0.8581～0.9318,模型检验的均方根误差(RMSE)、相对均方根误差(RRMSE)和相关系数(r)分别为0.1%～0.2%、4.0%～8.5%和0.9041～0.9854;基于RVI_CGMD的幂函数模型可较准确地预测LNA,模型R²为0.8684～0.9577,模型检验的RMSE、RRMSE和r分别为0.37～0.89 g·m^-2、6.7%～20.4%和0.9191～0.9851。与化学分析方法相比,利用CGMD可便捷准确地获取早、晚稻的LNC和LNA,在双季稻丰产高效栽培和氮肥精确管理中具有应用价值。

Monitoring leaf nitrogen concentration and nitrogen accumulation of double cropping rice based on crop growth monitoring and diagnosis apparatus

To verify the accuracy and adaptability of crop growth monitoring and diagnosis apparatus (CGMD) in monitoring nitrogen nutrition index of double cropping rice, we established a monitoring model of leaf nitrogen concentration (LNC) and leaf nitrogen accumulation (LNA) for double cropping rice based on CGMD. Eight early and late rice cultivars were selected and four nitrogen application rates were set up. The differential vegetation index (DVI), normalized difference vegetation index (NDVI) and ratio vegetation index (RVI) were collected using CGMD. Meanwhile, ASD FH2 high spectrometer was used to collect canopy spectral reflectance and calculated DVI, NDVI, and RVI. To verify the accuracy of CGMD, we compared the canopy vegetation indices change characteristics collected by CGMD and ASD FH2. The CGMD-based monitoring models of LNC and LNA were established, which was tested with independent field data. The results showed that LNC, LNA, DVI, NDVI and RVI of early and late rice increased with increasing nitrogen application rate, and increased first and then decreased with the advance of growth progress. The determination coefficient (R²) of fitting for DVI, NDVI and RVI from CGMD and ASD FH2 were 0.9350, 0.9436 and 0.9433, respectively. This result indicated that the measurement accuracy of CGMD was high, and that the CGMD could be used to replace ASD FH2 to measure canopy vegetation indices of early and late rice. Compared with the three canopy vegetation indices based on CGMD, the correlation between NDVI_CGMD and LNC and that between RVI_CGMD and LNA was the highest. The exponential model based on NDVI_CGMD could be used to accurate estimate LNC with the R² in the range of 0.8581-0.9318, and the root mean square error (RMSE), relation root mean square error (RRMSE) and correlation coefficient (r) of model validation in the range of 0.1%-0.2%, 4.0%-8.5%, and 0.9041-0.9854, respectively. The power function model based on RVI_CGMD could be used to estimate LNA with the R² in the range of 0.8684-0.9577, and the RMSE, RRMSE and r of model validation in the range of 0.37-0.89 g·m^-2, 6.7%-20.4% and 0.9191-0.9851, respectively. Compared with the chemical testing method, using the CGMD could conveniently and accurately measure LNC and LNA of early and late rice, which had a potential to be widely applied for high yield and high efficiency cultivation and precise management of nitrogen fertilizer in double cropping rice production.