冬小麦叶片花青素相对含量高光谱监测

为探究冬小麦叶片花青素含量的高光谱监测方法,以陕西省关中地区冬小麦为研究对象,分析了叶片光谱反射率与花青素含量的相关性,建立以不同波段组合的RSI、DSI和NDSI光谱指数为自变量的一元回归模型以及利用偏最小二乘法构建的多元回归模型,并进行模型精度比较。结果表明,所有模型中,开花期的PLS模型精度最高,预测效果最好(建模r~2=0.872 3,RMSE=0.005 9;检验r~2=0.912 8,RMSE=0.004 8),是预测冬小麦花青素的最优模型;各生育时期中,开花期模型精度较高,表现稳定,是预测冬小麦花青素的最佳生育时期。

Hyperspectral Monitoring of Anthocyanins Relative Content in Winter Wheat Leaves

In order to explore the hyper spectral estimation method of anthocyanin content in winter wheat leaves, and provide a theoretical basis for the efficient, non-destructive and large-area monitoring of anthocyanin content in winter wheat leaves,the winter wheat in Guanzhong area of Shaanxi province was used as the research material, and the correlation between spectral reflectance and anthocyanin content was analyzed. The regression model of RSI, DSI and NDSI spectral index with different band combinations was established. The multivariate regression model constructed by partial least squares(PLS)method is used to compare the accuracy of the model.The results showed that the PLS model at flowering stage had the highest precision and the best prediction effect (Modeling:r= 0.872 3, RMSE=0.005 9; Testing:r=0.912 8, RMSE=0.004 8), which is the optimal model for predicting winter wheat anthocyanins. During the growth period, the models at the flowering stage have higher precision and stable performance, which is the best growth period for predicting winter wheat anthocyanins. The research results can provide scientific basis and effective means for rapid acquisition and growth monitoring of winter wheat farmland information.