无人机载多光谱遥感监测冬油菜氮素营养研究

为探索无人机搭载的多光谱相机对冬油菜冠层氮素营养状况监测的可行性，设置9种施氮水平的油菜试验小区，获取八叶期、十叶期、十二叶期和蕾臺期的多光谱影像，同步采样分析获取地上部生物量、叶片氮浓度和氮素积累量等氮营养指标。以宽波段植被指数和氮营养指标的相关性为基础，通过敏感性分析确定最佳指数，建立预测模型并进行精度验证。结果显示，宽波段植被指数与氮营养指标有极显著的相关性，不同生育期差异明显。其中，红光标准值和蓝光标准值在蕾臺期均与各氮营养指标相关关系最好，且敏感性因子的值小而稳定。进一步研究表明，三种指标均可用红光标准值和蓝光标准值建立的二次模型进行估计，决定系数R2均大于0.85，模型精度较高，说明无人机多光谱遥感能有效辅助冬油菜氮素营养监测。

Monitoring of nitrogen nutrition in winter rapeseed using UAV-borne multispectral data

To use the filter-type multi-spectral camera mounted on the unmanned aerial vehicle (UAV) to dynamically monitor the nitrogen (N) nutrition of winter rapeseed, field experiment was conducted under 9 N application rates. The multi-spectral images of winter rapeseed canopy were obtained. Meanwhile, the conventional N monitoring parameters of aboveground biomass, leaf N concentration and N accumulation at 8-leaf period, 10-leaf period, 12-leaf period and bud period were taken. Based on the correlation between wide-band vegetation indices and nitrogen nutrition indices, the optimal vegetation indices were determined by sensitivity analysis. The models of nitrogen nutrition indices were calibrated and tested respectively, using training and test dataset. Results showed that broad-band vegetation indices had a significant correlation with N nutrition indices at each growth stage. Among them, normalized redness intensity (NRI1) and normalized blueness intensity (NBI) had the best correlation with N nutrition indices in bud stage. Small and stable noise equivalent (NE) values indicated the best vegetation indices for dynamic diagnosis of N nutrition in this crop. Further studies showed that the N concentration of leaves, aboveground biomass and nitrogen accumulation at bud stage could be estimated by quadratic model established by NRI1 and NBI with high accuracy. The study demonstrated multi-spectral remote sensing of UAV, which was flexible and promising in N nutrition monitoring of winter rapeseed.