基于RS与GIS的土地覆被变化及生态环境评价研究

发布时间：2018-03-03 05:20

本文选题：RS与GIS技术　切入点：土地覆被　出处：《江西理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：目前,制约人类生存和发展的问题有很多,但是土地覆被变化与生态环境恶化是人类面临的最紧迫问题。环保是人类为改善生存环境达成的唯一共识。除此之外,当前环境问题日趋严峻,在此形势下,关于土地覆被变化以及生态环境两方面的研究便成为当前阶段的研究热点与主要方向。分析土地覆被变化和生态环境变化规律,并对这些内容进行评价,能够为保护和治理与人类息息相关的生态环境提供有力的帮助。以土地利用以及生态环境两方面的理论为研究基础,以万安县为实际案例展开探究,数据源自高分一号遥感卫星,结合获取的数据,展开传统的分类方法以及面向对象分类方法的探究,研究两种方式是否适合高分影像土地覆被信息的提取;以2013年高分一号遥感影像数据为参照,对2013-2016年万安县用地变化进行了动态变化研究;以2016年Landsat8遥感影像及万安县统计年鉴作为研究区生态环境评价因子提取的数据源,使用层次分析法,并将提取生态环境评价因子作为前提,从而构建出生态环境评价模型,对研究区的生态环境进行综合且全面的评价。文章中的研究内容为:通过国产高分一号卫星获取研究区2016年遥感影像,同时运用非监督、监督及面向对像分类等方式将影像区分分类。利用验证样本进行分类结果验证得到非监督分类、监督分类、面向对象分类的混淆矩阵,通过混淆矩阵对比三种分类方法的总体精度与Kappa系数,结果显示运用传统分类方法展开高分影像土地覆被信息的提取是存在限制的。总体说来,在准确性方面,运用监督分类相比非监督分类是存在优势的,在高分影像的分类上,不论是监督分类还是非监督分类都无法达到面向对象分类的效果;将2013年以及2016年两个时相的高分一号遥感影像解译结果相互对比,能够知晓土地覆被变化的状况,即林地以及农用地呈现显著的下降状态。将2016年Landsat8影像当作数据源,从中提取例如归一化差异水体指数、土壤湿度指数、归一化植被指数和归一化建筑指数等生态环境评价因子;结合研究区统计年鉴及其他专题数据,获得数字高程模型、坡度、坡向、人口密度、地区生产总值等生态环境评价因子。从地形条件、自然条件、社会经济条件等3个层面9个指标出发,建立较为完善的万安县生态环境评价体系,结合主观权重与客观权重的层次分析方法,有层次的对每个指标分别赋权,并且把这些指标数据标准化,利用Arc GIS进行栅格计算及等级划分,得到万安县生态环境总体情况,发现万安县生态环境优、良等级占绝大部分,但也存在差和较差的区域。综合各阶段的研究成果,结合RS与GIS对高分一号遥感影像使用不同的分类方法提取其土地覆被信息,讨论高分影像在城市土地利用监测中的适宜性,扩展高分数据的运用领域;对不同的生态环境因子使用不同的提取手段,在评价因子的基础上建立评价模型,其后对生态环境进行评估是展开生态管理与治理不可或缺的。
[Abstract]:At present, there are many restricts the survival and development of mankind, but the land cover change and deterioration of the ecological environment is the most urgent problem facing humanity. Environmental protection is the only consensus to improve the human living environment is achieved. In addition, the increasingly serious environmental problems, in this situation, the research on land cover change and two the ecological environment has become a research hotspot in current stage and the main direction. Analysis of land cover change and ecological environment, and to evaluate the content, which is closely related to ecological environment for the protection and administration of human help. In two aspects of land use and ecological environment for the study of the theory basis to vie County as the actual case inquiry, data from one score satellite remote sensing data, combined with the expansion, the traditional classification method and object oriented classification The method of inquiry, of two is suitable for extracting high-resolution images of land cover information; by 2013 one score of remote sensing image data as a reference, 2013-2016 years of Wanan land were studied by dynamic change; statistics in 2016 Landsat8 remote sensing image and Wanan Yearbook as extraction of the ecological environment of the study area evaluation factor data the source, using the analytic hierarchy process, and the extraction of ecological environment evaluation factors as the premise, in order to construct the evaluation model of ecological environment, comprehensive and comprehensive evaluation on the ecological environment of the study area. The research content in the study area in 2016: access to remote sensing images by means of domestic a high score satellite, while the use of non supervision, supervision and like for the classification of image classification. The validation samples distinguish the classification results verify the unsupervised classification, supervised classification, object oriented classification The confusion matrix, through the overall accuracy and Kappa coefficient of confusion matrix comparison of three classification methods. The results show that using the traditional classification method of land cover information extraction of high-resolution images is limited. Generally speaking, in terms of accuracy, using supervised classification compared to unsupervised classification is of advantage, in the classification of the image score on whether supervised classification or unsupervised classification can not meet the object-oriented classification results; 2013 and 2016 two phase high one of remote sensing image interpretation results of the comparison, to know the land cover change, forest land and agricultural land showed a significant decline. The Landsat8 image in 2016 as the data extracted from the source, such as normalized difference water index, soil moisture index, evaluation of ecological environment construction and normalized vegetation index index node; Research area statistical yearbook and other special data, obtain the digital elevation model, slope, slope, population density, evaluation of ecological environment GDP. From the topography, natural conditions, social and economic conditions of the 3 level 9 indicators, the establishment of ecological environment Wanan perfect assessment system, combined with the subjective the weight and the objective weight AHP method, hierarchical weighted respectively for each index, and the index of data standardization, grid calculation and classification using Arc GIS, get the overall situation of ecological environment of Wan'an County, found that the ecological environment of Wanan County, good grades accounted for the vast majority, but there are poor and poor the regional research results. In each stage, the combination of RS and GIS on the high one remote sensing image using different classification methods to extract the information of land cover, discuss the high image in city land It is necessary to expand the application field of high score data by monitoring the suitability, use different extraction methods to different ecological environmental factors, and establish evaluation models based on the evaluation factors. Subsequently, the evaluation of the ecological environment is indispensable for ecological management and governance.

【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P237;P208;X826

【参考文献】