基于粒子滤波和多变量权重的冬小麦估产研究

发布时间：2018-01-14 14:13

  本文关键词：基于粒子滤波和多变量权重的冬小麦估产研究　出处：《农业机械学报》2017年10期 　论文类型：期刊论文

  更多相关文章： 冬小麦 粒子滤波 数据同化 遥感 熵值法 单产估测

【摘要】：为了构建能够反映作物长势的综合性指标以及准确估测作物产量,采用粒子滤波算法同化CERES-Wheat模型模拟和基于Landsat数据反演的叶面积指数(Leaf area index,LAI)、地上生物量和0~20 cm土壤含水率,获取冬小麦主要生育期以天为尺度的变量同化值,分析不同生育时期的LAI、地上生物量和土壤含水率同化值与实测单产的相关性,并应用熵值的组合预测方法确定不同状态变量影响籽粒产量的权重,进而生成综合性指数,并分析其与实测单产的相关性。结果表明,LAI、地上生物量和土壤含水率同化值和田间实测值间的均方根误差(Root mean square error,RMSE)以及平均相对误差(Mean relative error,MRE)均低于这些变量模拟值和实测值间的RMSE和MRE,说明数据同化方法提高了时间序列LAI、地上生物量和土壤含水率的模拟精度。基于不同状态变量的权重生成的综合性指数与实测单产间的相关性大于单个变量与实测单产间的相关性;基于综合性指数构建小麦单产估测模型,其估产精度(R2=0.78,RMSE为330 kg/hm2)分别比基于LAI、地上生物量和土壤含水率建立模型的估产精度显著提高,表明构建的综合性指数充分结合了不同变量在作物估产方面的优势,可用于高精度的冬小麦单产估测。
[Abstract]:In order to construct a comprehensive index which can reflect crop growth and estimate crop yield accurately. The particle filter algorithm was used to assimilate the CERES-Wheat model and the leaf area index leaf area index based on Landsat data. The aboveground biomass and soil moisture content of 0 ~ 20 cm were used to obtain the variable assimilation value of winter wheat in the main growth period, and to analyze the LAI of different growing stages. The correlation between the assimilation value of aboveground biomass and soil moisture content and the measured yield, and the combined prediction method of entropy value was used to determine the weight of different state variables affecting grain yield, and then the comprehensive index was generated. The correlation between the yield and the measured yield was analyzed. Root mean square error is the root mean square error between the assimilation value of aboveground biomass and soil moisture content and the measured value in the field. RMSE) and the mean relative error (mean relative error) are lower than the RMSE and MRE between the simulated and measured values of these variables. The data assimilation method improves the time series LAI. Simulation accuracy of aboveground biomass and soil moisture content. The correlation between the comprehensive index based on the weight of different state variables and the measured yield is greater than that between the single variable and the measured yield. The yield estimation accuracy of wheat yield estimation model based on comprehensive index was 330kg / hm ~ (-2), which was 330kg / hm ~ (2) than that based on LAI, respectively. The precision of yield estimation of aboveground biomass and soil moisture content model was significantly improved, which indicated that the constructed comprehensive index fully combined the advantages of different variables in crop yield estimation, and could be used to estimate yield per unit yield of winter wheat with high precision.
【作者单位】： 中国农业大学信息与电气工程学院;农业部农业灾害遥感重点实验室;陕西省气象局;
【基金】：国家自然科学基金项目(41371390)
【分类号】：S512.11;TP79
【正文快照】： 引言小麦是我国重要的粮食作物之一,其产量95%以上源于光合作用,而地上生物量是小麦光合作用的最终产物,与籽粒产量形成密切相关,因此,区域尺度小麦地上生物量的估算能够为籽粒产量的估测和预测提供重要依据。随着空间信息技术的发展,利用遥感技术获取地表植被信息和相关参数,，

本文编号：1423928