黄河三角洲地区土地资源丰富，开发前景广阔，但土壤盐渍化严重，影响着当地的农业生产，对生态环境的稳定性构成威胁。盐渍土地的改良对实现该区域农业和生态资源的可持续利用具有重要的意义，而改良盐渍土地的前提是掌握大面积的土地盐渍化程度。应用遥感技术，结合实测地物光谱和多源影像数据进行盐渍状况调查制图，为快速、准确、全面地监测盐渍化状况提供了可能。本文利用黄河三角洲地区CBERS-02B遥感影像数据，结合实测土壤全盐含量数据来反演该试验区域盐渍土可溶性全盐含量。针对盐渍土对应的遥感影像光谱反射率数据，采用相关性分析、诊断指数分析方法，确立了表征不同盐分盐渍土遥感影像反射率数据与盐分关系的敏感波段：band1（0.45-0.52μm），band2（0.52-0.59μm），band3（0.63-0.69μm）。将上述3个特征波段反射率数据作为输入，以盐渍土全盐含量数据作为输出，分别建立盐分反演的多元线性回归模型和BP人工神经网络模型，并对这两个模型进行精度检验。试验证明：基于CBERS-02B遥感影像特征波段反射率数据，利用建立的多元线性回归模型反演盐渍土含盐量，反演的精度较低，说明基于传统的多元线性回归模型反演土壤的盐分含量，其效果并不理想。利用BP神经网络很强的非线性映射能力，构建基于BP神经网络的土壤含盐量遥感反演模型，可以实现土壤含盐量的定量反演，为土壤含盐量的反演提供了一种新的方法和手段。并且该模型在高盐度区域（全盐含量>1%）的盐分反演中精度更高，其相对误差大部分在30%以内。通过构建的BP人工神经网络模型，结合CBERS-02B遥感影像将研究区盐渍土含盐量空间分布现状直观地显示出来，得到该地区的土壤含盐量结果图，并将研究区内不同等级盐渍土含盐量的差异直接区分出来。表明利用CBERS-02B遥感影像可以实现大范围内土壤含盐量的定量反演。该研究成果可以作为土地调查和农业生产等领域的参考资料，给有关部门在黄河三角洲盐渍化防治方面提供有益的决策支持。本文的研究有利于推动盐渍区域定量遥感研究的发展。; The Yellow River Delta has broad development prospects with rich land resource, but serious soil salinization affects local agricultural production and poses a threat to the stability of the ecological environment. Improvement and utilization of saline soil in the achievement of the regional agricultural and ecological sustainable use of resources is of great significance, while improving saline land is on the premise that the extent of land salinization has been known. Remote sensing technology is used in investigation and mapping about soil salinization with the measured spectrum data and multi-spectral image data. So it is possible to monitor salinization fast, accurately and comprehensively.The CBERS-02B’s remote sensing data of the Yellow River Delta and the measured soil salt content data were used to inverse the soluble total salt content of the saline soil in this test area. The sensitive bands: band1, band2, band3, which had better relations with the soil salt content data were determined by using the correlation analysis and the diagnostic index analysis.The salinity inversion models of multiple linear regression and BP artificial neural network were established with the three sensitive bands as input and the total salt content of saline soil as output. It could be found by testing the accuracy of the two models that the traditional multiple linear regression model couldn’t inverse saline soil salt content perfectly, with the CBERS-02B reflectance data of remote sensing images as input which could be verified by the low inversion precision.As BP artificial neural network model had strong non-linear mapping ability, this model could achieve the exact inversion of the saline soil salt content and provide a new method to inverse soil salt content. Especially, BP artificial neural network model had more accurate inversion result at high salinity area (total salt content >1%)，and the relative error of most zone were less than 30%.The soil salt content map was generated by running the trained BP artificial neural network model with the CBERS-02B remote sensing reflectance data as input. In this map, the spatial distribution of saline soil salt content of study area could be displayed visually. That was to say, the exact inversion of soil salinity in a wide range could be achieved with CBERS-02B remote sensing images.The research result of this paper can be acted as reference material of land investigation and agricultural production, which offers the related official department helpful decision-making support in reducing and controlling salinization in the Yellow River Delta. This study will promote quantitative development of the research on regional salinization by means of remote sensing.
修改评论