基于高光谱数据的新疆区域土壤盐渍化遥感应用机理和定量监测研究

Soil salinization has become a global issue, regional soil salinity monitoring using hyperspectral remote sensing is one of the foci and frontier of soil salinization research in recent years. However, the current study shows that the soil salinitation monitoring using remote sensing technique had a strong uncertainty, a weaker quantification and generality in different regions, especially, the sensitive spectral bands on soil salinity were significantly different, the data processing effect for improving the information extraction accuracy of soil salinity using remote sensing were very different, and the mechanism that cause these problems were unknown currently. In addition, that the soil salinity monitoring using remote sensing technique confined to surface soil layer cannot meet the needs of regional soil salinization control at present. In view of this, this project explores to use spectral decomposition method to study the relationship between the soil salinity, soil salt components, as well as other factors of soil and soil hyperspectral in lab, to monitor soil salinitation quantitatively using hyperspectral remote sensing images, combined the study of soil salinity monitoring using remote sensing with it of soil salinity change mechanism, to clarify the application mechanism of soil salinization quantitative monitoring using remote sensing technique based on above researches, and to develop soil salinity quantitative monitoring methods that are suitable for different regions. Meanwhile, with the aid of ground detection technology of profile soil salinity, by analyzing the relationship between soil salinity in soil layers with different depths, the deeper soil layer salinity can be detected using optical remote sensing. Then, the soil salinization will be detected using hyperspectral remote sensing images in Xinjiang province.

土壤盐渍化已成为全球性问题，区域土壤盐分高光谱遥感监测是近年土壤盐渍化研究的前沿和热点之一。目前土壤盐渍化遥感监测不确定性强、定量化和通用性较差，突出表现为，在不同区域，反映土壤盐分的光谱特征波段明显不同，为提高反演精度的遥感数据处理方法对改善土壤盐分信息提取精度的效果差别很大，且这些问题产生的机理不清楚。另外，目前局限于表层土壤盐分的遥感监测不能满足区域土壤盐渍化防治的需要。鉴于此，本项目探索利用混合光谱分解等方法，进行土壤盐类及其组成等和土壤光谱关系的实验室分析以及基于高光谱数据的土壤盐分遥感定量监测研究，将土壤盐分遥感监测与土壤盐分变化机理研究结合，阐明土壤盐渍化遥感定量监测的机理，发展适合于不同区域的土壤盐分遥感定量监测方法。同时，以剖面土壤盐分地面探测技术为辅助，通过分析不同土层盐分的关系，实现基于光学遥感的深层土壤盐分监测。在以上基础上，实现新疆区域土壤盐渍化的遥感定量监测。

土壤盐渍化已成为全球性问题，区域土壤盐分高光谱遥感监测是近年土壤盐渍化研究的前沿。目前在不同区域，反映土壤盐分的光谱特征波段明显不同，提高反演精度的遥感数据处理方法对改善土壤盐分信息提取精度的效果差别很大，且这些问题产生的机理不清楚。针对以上问题，本项目取得的创新和贡献如下：.对土壤盐渍化光学遥感定量监测的创新和贡献是：发现和验证了混合盐类土壤中，如果土壤主要盐类组成比较单一，尤其是只有一种主要盐类情况下，在一定的土壤盐分含量下，土壤盐分和土壤主要盐成分含量光学遥感监测敏感波段具有一致性特征，在这种情况下，土壤盐分和土壤光谱反射率的关系受主要盐类控制，光学遥感在在一定程度上可以用来分析土壤主要盐类组成情况；发现和验证了土壤盐分和土壤光谱反射率一般为线性或近似线性特征，在两种盐类组成的混合盐类中，如果两种盐质量比例差别不大或者质量占比相等，土壤盐分和土壤光谱反射率的关系可能具有显著的非线性关系特征，在土壤盐渍化光学遥感监测中要考虑非线性关系对土壤盐分估算精度的影响；证明了影响土壤盐分监测敏感波段选择的因素有土壤盐种类、土壤类型、土壤含盐量以及混合盐中各盐类数量和质量比例等；在某些土壤中，某些盐类在土壤中的含量和土壤光谱反射率的相关关系，在某些波段，会随着土壤中该类盐含量的增加而减弱；在土壤盐渍化光学遥感监测研究领域上，将目前主要是土壤盐分和土壤光谱之间关系的研究扩展到土壤盐分、土壤盐成分和土壤光谱之间关系的研究，研究领域更微观、更精确。.对区域土壤盐渍化光学遥感监测的创新和贡献是：分析得到研究区不同深度土壤盐分呈幂函数关系，在区域土壤盐分上升期，存在利用光学遥感间接监测深层土壤盐分的可能性；提出了集成土壤盐渍化地学机理与地表参数的区域土壤盐渍化光学遥感监测建模方法，发现深层土壤含水量、植被信息等能够提高研究区表层土壤盐分估算精度。