基于极化SAR的沙地土壤含水量反演方法研究

发布时间：2018-05-07 11:36

本文选题：极化SAR + 沙地　；参考：《内蒙古工业大学》2017年硕士论文

【摘要】：内蒙古草原是我国北方重要的生态屏障,生态环境极为脆弱。由于人类活动和气候变化的影响,导致内蒙古草原荒漠化越来越严重。草原荒漠化不仅会对我国的生态系统产生负面作用,而且也会对我国北方的经济发展产生影响。所以,对内蒙古草原荒漠化进行监测是十分必要的。内蒙古浑善达克沙地是监测内蒙古草原荒漠化的典型区域,土壤含水量是监测草原荒漠化的重要因素之一。因此,研究利用极化SAR反演大尺度内的沙地土壤含水量,对内蒙古草原荒漠化的监测具有重要意义。本论文的主要研究成果包括以下几个方面:(1)针对基于H-α和Wishart的分类方法存在的沙地地物与散射机理不对应的问题,研究了一种基于决策树和Wishart分类器的极化SAR图像分类方法。使用该方法对浑善达克沙地研究区的极化SAR数据进行分类,分类结果说明了沙地地物的分类精度得到提高。(2)在分析Dubois模型的基础上,结合AIEM对自然地表的后向散射特性的仿真,研究了一种反演沙地土壤含水量的经验模型。该模型不需要考虑地表粗糙度,仅利用VV和HH极化的后向散射系数即可反演沙地土壤含水量,该模型解决了Dubois模型反演沙地土壤含水量过高的问题。使用该模型对来自浑善达克沙地极化SAR数据进行处理,反演结果证明了该模型的有效性。最后,结合本文研究的基于决策树和Wishart分类器的分类方法与本文研究的沙地土壤含水量经验模型,发展了一种沙地土壤含水量反演算法。
[Abstract]:Inner Mongolia grassland is an important ecological barrier in the north of China, and the ecological environment is extremely fragile. Due to human activities and the impact of climate change, desertification in Inner Mongolia grassland is becoming more and more serious. Grassland desertification not only has a negative effect on the ecosystem of China, but also has an impact on the economic development of northern China. Therefore, it is necessary to monitor grassland desertification in Inner Mongolia. Hunshandake sandy land in Inner Mongolia is a typical area for monitoring grassland desertification in Inner Mongolia, and soil water content is one of the important factors in monitoring grassland desertification. Therefore, it is of great significance to study the application of polarimetric SAR to the estimation of soil moisture content in sandy land in Inner Mongolia. The main research results of this paper include the following aspects: 1) aiming at the problem that the sand objects do not correspond to scattering mechanism in the classification methods based on H- 伪 and Wishart, a polarimetric SAR image classification method based on decision tree and Wishart classifier is studied. This method is used to classify the polarized SAR data of Hunshandak sandy land. The classification results show that the classification accuracy of sandy land objects has been improved. (2) based on the analysis of Dubois model, combining with the simulation of backscattering characteristics of natural surface by AIEM, An empirical model for retrieving soil moisture in sandy land was studied. The model does not need to consider the surface roughness and can only use the backscattering coefficients of VV and HH polarization to estimate the soil moisture content of sandy land. The model solves the problem of high soil moisture content in sandy land by Dubois model. The model is used to process the polarized SAR data from Hunshandak sandy land. The inversion results show that the model is effective. Finally, combining the classification method based on decision tree and Wishart classifier and the empirical model of soil moisture in sandy land studied in this paper, an inverse algorithm of soil moisture content in sandy land is developed.
【学位授予单位】：内蒙古工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S812.2;TN958

【参考文献】