盐渍化土壤水分微波雷达反演与验证

发布时间：2018-05-10 13:05

本文选题：土壤水分 + 遥感　；参考：《农业工程学报》2017年11期

【摘要】：土壤介电常数是微波遥感进行土壤含水率测量的物理基础,尤其介电常数实部是必须解决的问题,土壤介电特性的研究显得尤为重要。该文目的是试验与评价C波段RADARSAT-2 SAR(synthetic aperture radar)数据模拟土壤介电特性,进而反演土壤水分的性能。以受盐渍化影响较严重的内蒙古河套灌区解放闸灌域为试验区,首先回归分析了介电常数实部与SAR四极化后向散射系数、地表粗糙度的复杂关系,并与Oh经验模型对照,其决定系数R2为0.859 7,模拟精度较高;然后验证常用的2个介电常数模型,Dobson半经验模型、Hallikainen简化实部经验模型模拟的介电常数实部与实测值的决定系数R~2分别为0.935 9、0.869,表明2个模型均能模拟地表土壤水分与介电常数实部的密切关系;最后构建了Dobson模型、Hallikainen简化实部模型反演土壤含水率的模型,并与统计回归模型比照,其模拟数值与土壤实测值的决定系数R2分别为0.803 8、0.737 4、0.842 1,均方根误差RMSE分别为5.2%、5.7%、5%。Dobson模型与统计回归模型反演结果与实地土壤墒情分布较为吻合,具有良好的精度和适用性,从而建立了一个较为完整的土壤介电特性研究体系,为微波遥感监测土壤水分奠定了基础。
[Abstract]:Soil dielectric constant is the physical basis of soil moisture measurement by microwave remote sensing, especially the real part of the dielectric constant is a problem that must be solved, the study of soil dielectric properties is particularly important. The purpose of this paper is to test and evaluate C band RADARSAT-2 SAR(synthetic aperture raster data to simulate the dielectric properties of soil and then to invert the properties of soil moisture. Taking the irrigation area of Jiefang sluice in Hetao Irrigation District, Inner Mongolia, which is seriously affected by salinization, as the experimental area, the complex relationship between the real part of dielectric constant and the backscatter coefficient of SAR and the surface roughness is analyzed by regression analysis, and compared with the Oh empirical model. The coefficient of determination R2 is 0.859, and the simulation accuracy is high. Then, two commonly used dielectric constant models, Dobson semi-empirical model and Hallikainen simplified real part empirical model, were verified. The determination coefficients of the real part of dielectric constant and the measured values were 0.935 ~ 0.869, respectively, which indicated that both models could simulate soil moisture on the surface. The close relation with the real part of dielectric constant; Finally, the model of Dobson model Hallikainen simplified real model for soil moisture retrieval was constructed and compared with the statistical regression model. The determination coefficient R ~ 2 of the simulated value and the soil measured value is 0.803 ~ (80.737) ~ (4) ~ (0.842), respectively, and the root mean square error (RMSE) is 5.22 ~ 5.7cm. The inversion results of the model and the statistical regression model are in good agreement with the distribution of soil moisture in the field, and have good accuracy and applicability. Thus, a relatively complete system of soil dielectric properties was established, which laid a foundation for microwave remote sensing monitoring of soil moisture.
【作者单位】：内蒙古农业大学水利与土木建筑工程学院;
【基金】：国家自然科学基金项目(51249007、51569018、51169016) 内蒙古自然科学基金项目(2013MS0609) “十三五”国家重点研发计划项目(216YFC0501301)
【分类号】：S152.7

【参考文献】