基于多传感器数据融合的田间尺度土壤三维模拟

项目名称： 基于多传感器数据融合的田间尺度土壤三维模拟

项目编号： No.41471179

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 地质学

项目作者： 汪善勤

作者单位： 华中农业大学

项目金额： 85万元

中文摘要： 土壤三维模拟模型为研究田间尺度土壤物质循环和能量流动提供高分辨率的土壤信息和合理的建模参数，对于精细农业管理、生态环境评价和土壤资源保护等，具有十分重要的理论意义和应用价值。本项目拟研究田间尺度土壤属性三维模拟的理论和关键技术。主要探索可见/近红外光谱、X射线荧光光谱、电磁探测等土壤近端感测信息融合的理论与方法，优化多传感器联合采样方案以获得有效互补的观测数据；研究多传感器数据空间配准方法，建立以土壤属性精准估计为目的的多源多维数据的融合估计模型；研究土壤深度方程与二维空间插值方法的耦合机制，探索土壤属性的三维空间分布与可视化表达，通过模拟结果的不确定性分析，揭示农田土壤三维空间分异规律，为农田管理提供高分辨率土壤信息和科学依据。

中文关键词： 田间尺度；土壤属性；三维空间分析；多传感器数据融合

英文摘要： Soil three-dimensional simulation models provide high-resolution information and parameters for the studies on substance circulation and energy transformation in field-scale soil. They have played an important role in precision agriculture, soil protection, ecological environment evaluation, etc.. This project intends to explore the theories and key technologies of three-dimensional simulation of soil properties in field scale. The proximal soil sensing techniques, such as visible/near infrared spectra, X-ray fluorescence spectrum, electromagnetic detection will be used and evaluated to measure soil properties through the multi-source data fusion and optimized sampling scheme to obtain reasonablely complementary observation data. The data-fusion models of soil properites will be investigated to improve accuracy of soil properites prediction based on the spatial registration methods built for multi-sersor measurements. And then the depth function coupled with two dimensional Kriging will be explored by parameters transfer and derived on the observation profile sampls to set up the three-dimensional simulation model of soil properties. the visualization of verticle and horizontal distributions of soil propertiesin in study fields will be conducted, as well as the uncertainty assessment on models and predictions. The high-resolution spatial variances of soil in farmland are expected helpful to enhance the farm management and soil protection.

英文关键词： field scale;soil properties;three-dimensional spatial analysis;multi-sensor data fusion