面向农田重金属监测的空间插值方法研究【毕业论文】.docVIP

分类号 TP391 密级 公开 UDC 全日制工程硕士学位论文 面向农田重金属监测的空间插值方法研究 学 位 申 请 人： 专 业 领 域： 计算机技术 指 导 教 师： 教授 二○ 年四月 A Dissertation Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Computer Technology Spatial Interpolation Method Survy Based on Farmland Heavy Metal Monitoring Graduate Student: Major: Computer Technology Supervisor: Prof. China Three Gorges University Yichang, 443002, P. R. China April, 20 三峡大学学位论文声明 内 容 摘 要 土壤中重金属来源分析 克里格方法不同于其他空间插值法的地方在于它从变量自身的特点出发，考虑观测点的整体空间分布情况，可以对插值误差做出理论估计。它是基于采样数据反映区域化变量的结构信息。变差函数和协方差函数将临近事物比远处事物更相似这一假设加以量化。变差函数和协方差都将统计相关性的强度作为距离函数来测量。对变差函数和协方差函数数据建模的过程将变差函数或协方差曲线与经验数据拟合。目标是达到最佳拟合，并将现象的认知纳入模型。之后模型可用于预测。 针对常规克里格插值方法中协方差函数多是凭经验在有限确定的函数模型中选择的现状， 难以准确刻画实际数据分布( 特别是复杂要素资料的空间结构) 的不足， 基于最小二乘支持向量机从实际资料场中拟合重构协方差函数的研究思想，提出了一种改进的插值方法——支持向量机—克里格插值算法( LS-SVM-Kriging) 。Abstract Taking into account of the spatial correlation and heterogeneity, the spatial sampling has a higher efficiency and accuracy of the samples compared with classical sampling methods, therefore, it receives widespread attention. It is a trend to use multidimensional auxiliary variables to optimize the sampling locations at regional scale in the process of designing sampling method, and these variables have significant correlation with target variable. It is very important to study an efficient way to build an optimized feature space with the auxiliary variables. This paper proposes a novel method to build an optimized feature space with dimensionality reduction method. Select the appropriate data for a specific spatial interpolation method is very important, but not easy to do. The effect of spatial interpolation methods depends on many factors. What is the most appropriate method? it is not simply a question to answer. One way only because in certain situations is optimal. In making choices, many factors need to be considered. The choice of