基于RS和GIS的黄河三角洲盐碱地分级与治理研究

发布时间：2018-06-07 00:25

本文选题：黄河三角洲 + 盐碱地分级　；参考：《山东师范大学》2015年硕士论文

【摘要】：一直以来，盐碱地问题都是全球性的问题和难题。盐碱地破坏作物赖以生存的土壤环境，影响作物的正常生长，甚至导致作物死亡，阻碍了农业生产的正常进行和生态环境的平衡发展。据国际粮农组织及联合国教科文组织的相关统计，目前全球盐碱地的总面积约9.54亿hm2，其中我国所占面积为9900余万hm2。盐碱地问题已经成为制约我国农业经济发展的重要因素。黄河三角洲地区的盐碱地问题较为突出，已经严重制约该地区农业的正常发展，加强治理刻不容缓。所谓“知己知彼，百战不殆”，了解该地区的盐碱地分布状况与盐碱化程度是治理的首要任务。本文在黄河三角洲地区进行实地土壤采样，并对采样样本进行实验，测得采样点土壤样本的土壤含盐量。在ArcGIS10.1平台上生成采样点的点要素数据，并存储各采样点的土壤含盐量。对环境与灾害监测预报小卫星星座B星（HJ-1B星）的同期遥感影像进行预处理，并提取采样点位置各个波段的遥感影像反射率值。在SPSS平台上对影像各个波段反射率值与相应的土壤含盐量进行Pearson相关系数的运算，并计算各个波段的标准差，从而得到各个波段的诊断系数，以此来表明各个波段对土壤含盐量的敏感程度。结果表明，Band1、Band2和Band3波段的土壤含盐量敏感性较为统一，且明显高于Band4波段，此三波段可以用来进行土壤含盐量的遥感定量反演运算。在SPSS平台上，以Band1、Band2和Band3三个波段为自变量，使用逐步回归分析的方法进行多元线性回归分析运算，以建立影像反射率值与土壤含盐量之间的关系模型。结果显示，Band2波段自变量被剔除，Band1和Band3波段自变量作为保留自变量进行了回归运算，运算结果显著。 BP神经网络以其较强的非线性回归映射能力和突出的自学习自适应能力而被广泛使用。将BP神经网络引入到土壤含盐量的遥感反演中来，可以充分利用其自身特点，为遥感定量反演提供一种全新的方法和途径。本文在MATLAB平台上构建了一个包含输入层、隐含层和输出层的三层BP神经网络模型，使用2-16-1结构，即输入层2个节点，隐含层16个节点，输出层2个节点。将Band1和Band3作为输入节点，使用Log-Sigmoid作为训练函数，进行网络模型训练。通过精度检验对比BP神经网络模型和多元线性回归模型的预测能力，，结果显示，前者在大部分情况下要强于后者。因此，使用训练得到的神经网络模型对黄河三角洲地区的土壤含盐量进行仿真。结果表明，黄河三角洲地区的盐碱地盐碱化程度较为严重，其中以重度盐碱地和盐田为代表的原生盐碱地比重占到了近七成，且集中分布在黄河以及大片水域集中的地区附近。由于水势高、水量大等因素导致这些地区的地下水埋深变浅，从而引起盐碱地的盐碱化程度加深。最后，本文描述了黄河三角洲地区盐碱地的成因，并对治理提出了相关建议。尽管人为措施无法真正阻止该地区盐碱地的形成，但通过科学合理的方法可以达到缓解盐碱地盐碱化程度的目的，从而减轻盐碱地对农业发展的制约。
[Abstract]:The problem of saline - alkali land has always been a global problem and difficult problem . The saline - alkali land destroys the soil environment that crops depend on , affects the normal growth of crops , and has hindered the normal development of agricultural production and the balanced development of the ecological environment . According to the relevant statistics of FAO and UNESCO , the total area of the global salt - alkali land is about 900 million hm2 . The problem of saline - alkali land has become an important factor restricting the development of agricultural economy in China .

The problem of saline - alkali land in the Yellow River Delta region is more prominent , which has seriously restricted the normal development of agriculture in the region and strengthened its governance .

In this paper , the soil samples were sampled in the Yellow River Delta region , and the samples were tested to measure the soil salinity of soil samples . The point element data of the sampling points were generated on the ArcGIS 10.1 platform , and the soil salinity of each sampling point was stored .

The results show that the sensitivity of each wavelength band to soil salinity is higher than that of Band1 , Band2 and Band3 bands , and the three wavelength bands can be used for remote sensing quantitative inversion of soil salinity .

Based on the SPSS platform , using three bands of Band1 , Band2 and Band3 as independent variables , multiple linear regression analysis was performed using stepwise regression analysis to establish the relationship model between the image reflectivity value and the soil salinity . The results show that the Band2 band independent variable is eliminated , the Band1 and Band3 band independent variables are used as the reserved independent variables , and the operation result is remarkable .

The BP neural network is widely used in remote sensing inversion with strong nonlinear regression mapping ability and prominent self - adaptive ability . The BP neural network is introduced into the remote sensing inversion of soil salinity , and a new method and approach for remote sensing quantitative inversion can be fully utilized . In this paper , a three - layer BP neural network model including input layer , hidden layer and output layer can be fully utilized .

The results show that the salinity of the saline - alkali land in the Yellow River Delta region is more serious , and the proportion of the native saline - alkali land represented by the severe saline - alkali land and the salt field accounts for nearly 70 % , and the concentration of the salt - alkali land in the Yellow River is close to the area concentrated in the Yellow River and the large area .

Finally , this paper describes the genesis of saline - alkali land in the Yellow River Delta region , and puts forward some suggestions for the treatment . Although artificial measures cannot really prevent the formation of saline - alkali land in the region , the aim of relieving the salinity of saline - alkali land can be achieved by scientific and reasonable method , so as to reduce the restriction of saline - alkali land to agricultural development .
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：S156.4

【参考文献】