【作者】 徐涛；

【导师】 周廷刚；

【作者基本信息】 西南大学 ， 地图学与地理信息系统， 2010， 硕士

【摘要】 随着遥感技术的发展,高分辨率遥感影像在城市研究中得到广泛应用。高分辨率遥感影像由于其本身的优点为研究城市环境提供了重要的信息来源。利用高分辨率遥感影像,获取城市植被类型、分布及其结构,可以为改善城市环境质量、优化植被布局、改进城市绿地系统规划提供客观的现时依据,对于城市未来的绿化建设和总体规划具有重要意义。本文以重庆市北碚区为例,应用高空间分辨率QuickBird影像,采用面向对象的遥感解译方法,对城市地物进行分类,提取植被信息。建立基于高空间分辨率卫星影像的城市绿地解译模式,提高城市植被分类精度,定量分析城市植被数量、对城市植被进行评价和分析、为城市绿地规划相关决策部门提供客观依据。主要研究内容与研究成果如下：(1)影像预处理：通过定量分析各波段之间的相互关系,进行波段选择；几何校正；对高空间分辨率影像的全色波段与多光谱波段进行像素级融合。(2)利用面向对象的方法进行遥感解译：利用多尺度分割的方法,进行影像分割；构建解译地物的知识库；提取地物信息。(3)利用传统的分类方法(监督分类法与非监督分类法)进行解译,通过多种评价指标进行评价传统解译与面向对象解译方法的差异。(4)在遥感与GIS支持下,运用景观生态学原理,结合多种绿地评价指标对北碚区植被进行生态评价和景观分析,分析北碚区绿化现状、探讨北碚区城市绿地规划布局并提出合理性建议。更多还原

【Abstract】 With the development of remote sensing, high spatial resolution remote sensing images are extensively used in the studing of city. High spatial resolution remote sensing images can provide important information sources.To abtain the kinds, distribution and structure of urban vegetation by using high spatial resolution remote sensing images can provide objective foundation data. The results are very important to virescence construction and general planning.The paper takes the Beibei Chongqing for example. The aim is the information extraction of vegetation and the other things in the city by using the object-oriented method. The central contents of the paper are model struction, high classes precision and estimation and ananlysis about ubran vegetation, which are the foundations for government decision-making.The main harvests of studing:(1)Image pretreatment:the choice of available bands by quantitative analysis about all bands, Ortho-rectification, then by making use of high spatial resolution remote sensing images and pan image resolve images in the level of pix.(2)Oriented-object method which includes multi-scale segment, repository construction and information extraction is made use of extracting information about city.(3)The traditional classifications (supervised and unsuperviesed) are used extracton information. The methods are compared to the oriented-object method by many indexes.(4)On the bases of remote sensing and GIS, making use of landscape ecology theory and indexes about greenbelt estimation analysis greenbelt actuality and discuss the greenbelt distribution and provide reasonable suggestions.更多还原