基于高光谱技术的覆膜旱作水稻植株氮含量及籽粒产量估算

为建立利用光谱技术快速诊断覆膜旱作水稻植株氮营养和产量的估算模型,应用高光谱技术分析了长江中下游覆膜旱作区水稻拔节期和抽穗期内5种不同施氮水平(0、60、120、180和240kg/hm~2(N))下植株冠层光谱特征及其与植株氮素含量和产量的关系,并分别构建了植株含氮量和产量的估算模型。结果表明:拔节期和抽穗期内不同供氮水平下冠层光谱的变化规律基本一致,均随着供氮水平的增加,反射率在可见光区降低、在近红外区增大。覆膜旱作水稻植株氮含量与552和890nm 2个敏感波段构成的比值(RVI)和绿色归一化植被指数(GNDVI)的关系最佳。构建的水稻关键生育期植株全氮含量及水稻产量的估算模型预测效果均较好,其中植株全氮含量拟合方程的决定系数为0.730~0.808,采用拔节期的RVI对覆膜旱作水稻进行估产的决定系数达到0.724。本研究构建的模型可以用来估计该地区覆膜旱作水稻的氮素营养状况和作物产量。

Crop nitrogen content diagnosis and yield estimation in ground cover rice production system based on hyperspectral data

In order to establish the model for crop nutrient diagnosis and yield estimation in ground cover rice production system (GCRPS) by spectral technique, the hyperspectral technique was applied to explore the canopy spectral features at the jointing and heading stages in GCRPS, and the relationships between crop spectral indices and rice nitrogen (N) contents as well as grain yields at five different rates of N fertilizer (0, 60, 120, 180 and 240 (N) kg/hm²) were determined in the middle and lower reaches of the Yangtze River.The results showed the changing trends of the canopy spectral under different N fertilizer levels were similar at the jointing and heading stages, both showing a decreasing trend at visible band and an increasing trend at the near infrared waveband with the increasing of N fertilizer rate.The rice N contents had a close correlation with the ratio of vegetation index (RVI) and the green difference vegetation index (GNDVI) derived from the sensitive wavebands of 552 and 890 nm.The determinant coefficients (R²) of the regression equations for rice N contents ranged from 0.730 to 0.808 and the R² for rice yield estimation equation using RVI at the jointing stage reached to 0.724.We concluded that the established regression equations in this study can be used to estimate rice N contents at the key rice growth stages and grain yield in study area.