【摘要】 为探寻准确高效的冬小麦生物量动态监测方法,以2018—2020年SRS-NDVI观测仪监测数据为基础,将冬小麦生物量观测数据以返青期为界分为两个阶段。选取归一化植被指数（NDVI）、比值植被指数（RVI）、差值植被指数（DVI）、增强型植被指数（EVI2）,计算逐日累积植被指数（CVI）并分别进行曲线拟合分析,建立回归模型,研究各植被指数与实测冬小麦生物量之间的关系。结果表明:冬小麦播种至返青期,最优模型为二次多项式,基于NDVI的累积植被指数模型估测精度最高,y=-0.0479x²+7.0481x-25.5040,R²为0.9829,均方根误差（RMSE）为9.61,平均相对误差（MRE）为10.51%;冬小麦返青至成熟期,最优模型为幂函数,最佳估测模型仍为基于NDVI的累积植被指数模型,y=0.0126x^2.3938,R²为0.9553,RMSE为150.25,MRE为10.22%。因此,基于NDVI的累积植被指数是冬小麦生物量动态监测的最佳方法,可为作物自动化观测提供新的思路和方法。更多还原

【Abstract】 In order to explore the effective method for dynamic monitoring the winter wheat biomass, the biomass observation data of winter wheat were divided into two stages according to the regreening period based on the monitoring data of the SRS-NDVI observatory from 2018 to 2020, and normalized differential vegetation index（NDVI）, ratio vegetation index（RVI）, difference vegetation index（DVI）and enhanced vegetation index 2（EVI2） were selected to calculate the daily cumulative vegetation index（CVI） to study the relationship between vegetation and measured biomass of winter wheat by curve fitting analysis and establishing regression models. The results showed that from sowing to regreening stage, the optimal model was quadratic polynomial and the cumulative vegetation index model based on NDVI had the highest estimation accuracy, which was y=-0.0479 x²+7.0481 x-25.5040, R² was 0.9829, root mean square error（RMSE） was 9.61, mean relative error（MRE） was 10.51%. The optimal model was power function from regreening to ripening stage and the best estimation model was the cumulative vegetation index model based on NDVI, which was y=0.0126 x^2.3938 with the R² as 0.9553, RMSE as 150.25 and MRE as 10.22%. Therefore, the cumulative vegetation index based on NDVI was the best for dynamic monitoring of winter wheat biomass, and it could provide new ideas and methods for automatic observation of crops.更多还原