基于CA-Markov模型的杭州市土地利用变化研究

发布时间：2018-06-13 04:38

本文选题：GIS空间分析 + CA-Markov模型　；参考：《浙江师范大学》2015年硕士论文

【摘要】：当今中国正处于快速城市化阶段。城市用地的扩张带来了耕地减少、环境污染、生物多样性减少等一系列问题。研究城市土地利用变化,分析和预测城市用地扩展演化规律,是实现土地可持续利用的前提和基础。本文以杭州市城区作为研究对象,以元胞自动机和马尔科夫相关理论为主要方法模型,利用杭州城区1994年、2004年和2013年的遥感影像数据,通过ERDAS9.2软件对影像数据进行预处理和数据分类,利用ArcGIS9.3软件进行面积统计和数据转换,利用IDRISI17.0实现的CA-Markov模型研究杭州城区建设用地的发展演化规律,进行土地建设用地合理的预测。论文的主要工作包括以下方面：(1)土地利用类型数据处理。运用ERDAS对杭州市1994年、2004年、2013年的遥感影像统一坐标系为WGS_1984_UTM_Zone_50N,采用二次多项式纠正方法进行几何精校正。运用ArcGIS软件矢量化杭州市城区的行政边界图,并以行政边界图为掩膜,应用ArcToolbox工具箱对校正后的影像进行裁剪。再采用ERDAS人机交互的监督分类中最大似然法进行图像解译,提取出林地、耕地、园地、其他用地、居民点及工矿用地、水体六种类型的分类模板,对分类模板精度评价后生成六种土地利用类型的现状图。进行分类后处理以后,在ERDAS中将数据转换成Coverage格式,再由ArcGIS软件将Coverage格式转换成shp格式的1994年、2004年和2013年土地利用现状图。(2)土地利用变化分析。运用ArcGIS软件对杭州市林地、耕地、园地、其他用地、居民点及工矿用地、水体六种土地利用类型从数量、动态和空间方面进行分析,研究土地利用的变化趋势。(3)土地利用模拟研究。利用ArcGIS将shp数据转化成栅格数据,在IDRISI中对栅格数据进行RECLASS、DISTANCE和FUZZY处理,得到各类型的适宜性图。再由IDRISI中的Markov模块制作相应的土地利用转移面积矩阵。最后利用CA-Markov模型以2004年为基础图像,通过1994年到2004年的土地利用转移矩阵面积,和2004年的土地利用转变适宜性图集,来预测杭州市2013年土地利用现状,从数量和空间精度两个方面来验证2013年预测图和实际面积图的准确性,最终来验证CA-Markov模型预测的准确性。其中,最关键的是土地适宜性图集的制作。选取已建设用地、水体和基本农田作为约束因子,已建设用地、水体、交通线、城镇、坡度作为限制因子,分别制作林地、耕地、园地、其他用地、居民点及工矿用地、水体六种土地利用类型的适宜性图像,在Colletion Editor模块将六种土地利用类型的适宜性图像合成土地利用转变适宜性图集。(4)土地利用预测研究。以2013年为基础图像,通过2004年到2013年的土地利用转移面积矩阵,和2013年的土地利用转变适宜性图集,来预测杭州市2020年土地利用类型,为合理规划杭州市土地利用格局提供科学依据。
[Abstract]:Nowadays, China is in the stage of rapid urbanization. The expansion of urban land has brought about a series of problems, such as the decrease of cultivated land, environmental pollution and the decrease of biodiversity. It is the premise and foundation to study urban land use change and analyze and forecast the law of urban land expansion and evolution. In this paper, taking Hangzhou urban area as the research object, taking cellular automata and Markov correlation theory as the main method model, using the 1994, 2004 and 2013 remote sensing image data of Hangzhou urban area. Image data are preprocessed and classified by ERDAS9.2 software, area statistics and data conversion are carried out by ArcGIS9.3 software, CA-Markov model realized by IDRISI 17.0 is used to study the law of development and evolution of urban construction land in Hangzhou. Reasonable prediction of land construction land. The main work of this paper includes the following aspects: 1) Land use type data processing. The unified coordinate system of remote sensing images in Hangzhou in 1994, 2004 and 2013 is WGS1984S / s / s, and the quadratic polynomial correction method is used to correct the geometry of Hangzhou city by using ERDAS. Using ArcGIS software to vectorize the administrative boundary map of Hangzhou city, and taking the administrative boundary map as mask, the corrected image is clipped with ArcToolbox toolbox. Then the maximum likelihood method of ERDAS human-computer interaction was used to extract the classification templates of six types of forest land, cultivated land, garden land, other land, residential land and industrial and mining land, and water body. The present situation map of six land use types was generated after the accuracy evaluation of classification template. After classification and post-processing, the data were converted into coverage format in ERDAS, and then the coverage format was converted into shp format by ArcGIS software. Land use change was analyzed in 1994, 2004 and 2013. Using ArcGIS software, this paper analyzes six land use types of Hangzhou woodland, cultivated land, garden land, other land, residential land and industrial and mining land, and water body from the aspects of quantity, dynamics and space. To study the trend of land use change. The shp data is transformed into raster data by ArcGIS, and the raster data is processed by RECLASS DISTANCE and FUZZY in IDRISI. The corresponding land use transfer area matrix is made by Markov module in IDRISI. Finally, the CA-Markov model is used to predict the present situation of land use in Hangzhou in 2013 by using the land use transfer matrix area from 1994 to 2004 and the suitability map of land use transformation in 2004 by using CA-Markov model as the basic image. The accuracy of the 2013 prediction map and the actual area map is verified from the aspects of quantity and spatial accuracy. Finally, the accuracy of CA-Markov model is verified. Among them, the most important thing is the making of land suitability atlas. The construction land, water body and basic farmland are chosen as the constraint factors, and the construction land, water body, traffic line, town and slope are chosen as the limiting factors to make woodland, cultivated land, garden land, other land, residential land and industrial and mining land, respectively. In the Colletion Editor module, the suitability images of the six land use types were synthesized into the land use change suitability atlas. Based on the 2013 image, the land use transfer area matrix from 2004 to 2013 and the suitability map of land use change in 2013 were used to predict the land use types in Hangzhou in 2020. To provide scientific basis for rational planning of land use pattern in Hangzhou.
【学位授予单位】：浙江师范大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F301.24

【相似文献】