【作者】 周兴东；

【导师】 于胜文； 赵长胜；

【作者基本信息】 山东科技大学 ， 大地测量学与测量工程， 2010， 博士

【摘要】 利用遥感和地理信息系统等技术方法,积极开展城市化过程中的景观格局演变研究,寻求城市内部人地关系系统的协调与均衡发展,为徐州市的土地利用、城市建设和区域可持续发展提供有效决策支持,无疑具有重要的科学意义和现实意义。同时,在其他地区的类似研究中也可以得到参考和借鉴。本文以徐州市城区为研究对象,以1987年、1995年、2001年和2007年四个年份的美国陆地资源卫星Landsat TM/ETM+遥感影像数据为主要数据源,在对徐州市城区景观现状进行遥感影像分析和地面调查的基础上,以景观生态学等相关理论为依据,采用3S技术、景观分析方法,从格局、过程、驱动、演变、模拟全过程,研究了过去二十年徐州市城区景观格局的演变过程,建立了基于遥感和GIS监测、分析、评价、预测技术体系。具体研究内容如下：将研究区域的景观类型划分为：水体、建设用地、林地、其它植被、耕地、裸地六大类,分别采用最小距离分类法(MDC)、最大似然分类法(MLC)、支持向量机分类法(SVM)、决策树分类法(DTC)和面向对象的分类法(OOC),对研究区TM/ETM+遥感影像进行分类处理,对六种景观分别提出了其信息提取方法,同时认为研究区具有最好应用效果的支持向量机(SVM)分类方法,并且以这种方法分类的结果制成了各个时相的景观类型分布图。2、通过景观格局演变分析,揭示了研究区土地利用景观演变的特点利用景观格局分析软件Frgastatsversino3.3计算各种景观格局指标,研究了研究区1987-2007年景观格局演变的特征,并使用土地利用景观转移矩阵定量描述了1987-1995、1995-2001、2001-2007三个时段内各种景观类型相互之间的转化情况,重点研究了斑块类型水平上的景观演变和景观水平上的景观格局演变,得出了徐州市城区的土地利用景观演变的特点：耕地和建设用地在研究区占据主导地位。利用景观生态学相关理论,从粒度效应、幅度效应的角度,对20年来徐州市城区土地利用与景观演变的粒度效应、幅度效应进行了系统分析。研究发现：区域土地利用景观格局随粒度和幅度的变化出现显著变化,在对研究区景观格局进行分析时,需根据研究对象和目的的不同采用适合的研究尺度和景观格局指数。利用主成分分析法对土地利用景观演变的驱动机制进行了分析研究,结果认为：经济发展因子和人口因子是推动研究区土地利用景观演变的主导性驱动因素。针对城市景观格局模拟与预测的要求,采用Markov模型和CA-Markov模拟模型对研究区景观进行模拟预测的案例研究,实验证明CA-Markov模拟模型的模拟效果明显优于Markov模拟模型的模拟效果,利用CA-Markov模拟模型预测了研究区2013年的景观格局演变,分别得出了各种景观的演变情况。更多还原

【Abstract】 Based on the remote sensing and GIS technique, it studies the landscape pattern evolution in the process of urbanlization. It seeks harmonious and balanced development for the relationship between human beings and land in the urban inner. It can provide effective precision support for land use, urban constructing and regional sustainable development for Xuzhou city. So it has important scientific and practical meanings. And it will provide effective reference for similar study in other region.This paper chooses Xuzhou urban area as research object and selects the Landsat TM/ETM+ remote sensing image acquired in 1987,1995,2001 and 2007 as major data source. On the basis of remote sensing image analysis and ground survey for the landscape present situation in Xuzhou urban area, it studies the landscape pattern evolution and it builds the technical system for monitoring, analzing, evaluating, predicting based on the theories about landscape ecology,3S technique, and.the landscape analysis method.(1) it puts forward to the classification system of landscape, and studies the extracting methods for landscape information in Xuzhou urban area, then it puts forwards to some classification methods suitable for research areaIt divides the landscape type in the study area into six classes such as water, construction land, woodland, farmland, bare land and other vegetations. And it carries out classification processing based on MDC, MLC, SVM, DTC, OOC methods for the Landsat TM/ETM+ remote sensing images. Through the precision evaluation and contrast analysis for different classification methods, it puts forward effective classification method for 6 different landscapes respectively. And it finds that SVM method is suitable for study area. Then it maks landscape type distribution map for every period based on SVM method.(2) it reveals the characteristics of land use landscape evolution through analyzing the landscape pattern evolutionIt computes some landscape pattern indices using the soft Frgastatsversino3.3. It studies the characteristics of landscape pattern evolution from 1987 to 2007. And it quantitatively describes the transformation cases of every landscape type from 1987 to 1995, from 1995 to 2001, and from 2001 to 2007 using landscape transformation matrix. It studies the landscape pattern evolution in class level and in landscape level. It gets the characteristics of landscape evolution in Xuzhou urban area. The farmland and construction land play a leading role.(3)it studies the scale effects about landscape index, it can provide the basis for selecting the granularity and amplitudeBased on corresponding theories about landscape ecology, it analyzes the granularity effect and amplitude effect of the land use and landscape evolution about Xuzhou urban area for 20 years. It finds that the land use landscape pattern will change obviously with the change of granularity effect and amplitude effect of the landscape pattern of regional land use. When analyzing the landscape pattern of study area, it needs to adopt suitable study scale and landscape pattern index according to different object and objective.(4)it studies the land use landscape evolution driving mechanism, and it reveals the dominating factors promoting the land use landscape evolutionThe driving mechanism for land use landscape evolution has been studied using.principal component analysis. The conclusion is that the economic development factor and population factor are two dominating driving factors.(5)it puts forward to the prediction model for landscape pattern evolution, and it predicts the landscape pattern in 2013According to the demand for urban landscape pattern simulation and forecast, it carries out the case study of landscape pattern simulation and forecast adopted Markov model and CA-Markov model. The test proved that CA-Markov model is better than Markov model in simulation effect. Then it forecasts the landscape pattern evolution of study area in 2013 year based on CA-Markov model. It gets every landscape type evolution individual.更多还原