土地利用/土地覆被变化的遥感监测及空间动态预测

【作者】 曲凯；

【导师】 李新运；

【作者基本信息】 山东师范大学 ， 地图学与地理信息系统， 2007， 硕士

【摘要】 土地利用/土地覆被变化(LUCC)反映了一个区域的社会及生态环境的变化情况,随着经济发展和城市化进程的加快,引起了土地利用及土地覆被的剧烈变化,全面及时地掌握资源环境的基本状况就显得至关重要。采用先进的遥感技术和GIS技术对土地利用变化进行快速的动态监测,可以及时准确地掌握土地利用的空间信息和时间序列信息,利用基于复杂系统思想的细胞自动机模型(Cellular Automata,以下简称CA)研究LUCC过程的复杂行为,定量探讨LUCC过程并进行未来土地利用状况的预测,可为区域决策者进行土地利用评价提供依据。因此,选择较小空间范围的典型地区开展LUCC过程的遥感监测和空间动态预测研究,不仅可以为地方经济发展提供决策支持,而且对丰富全球土地利用变化研究具有重要意义。本文以济南市市辖区为研究区,以格局、过程、机理和模型研究为核心,利用两个时相的遥感图像对研究区2000年至2004年间的土地利用/土地覆被变化进行了动态监测,并运用基于CA的空间模型对未来土地利用演变进行了模拟与预测,并用基于信息熵理论的离散空间场相关指数对预测结果进行检验。论文的主要研究内容和研究成果有:(1)用遥感图像进行土地利用/土地覆被变化的监测是一种快速有效的方法。本研究运用RS和GIS技术,在对每期图像进行预处理和增强处理的基础上,采用监督分类结合目视判读的提取方法,实现了研究区土地利用信息较高精度的提取。监测结果表明:济南市在2000-2004年间,土地利用/土地覆被整体变化不大,只有城乡工矿居民用地和未利用土地发生了较大的变化;济南市总体的土地利用程度正向集约用地型转变;在发生变化的类型中,耕地因区位条件较优和比较经济效益的差异而不断向其他用地类型流转,是其它土地类型新增面积的主要来源,未利用地大多分布滩地附近,主要开发为耕地。(2)通过定义局部细胞邻域关系以及使用合适的局部转换规则,CA可以模拟和表示整个系统中复杂现象的时空动态变化,非常适用于具有复杂时空特征的地理系统的研究。本研究通过建立基于CA的土地利用空间动态预测模型,对2008、2012年两个时间段的土地利用演变进行了模拟与预测,从预测结果看,随着时间的推移,济南市的耕地面积不断减少,城乡工矿居民用地面积不断增加,并且沿着城区的边界增长,而林地、草地、水域和未利用地的变化则不大。(3)通过CA模拟与预测的土地利用分布数据、城市扩展数据都是空间上连续而属性状态离散的离散空间场,空间统计学中各种相关指数和灰色系统理论中的关联度指数不能计算其相关性,本研究基于信息熵理论,构造一个具有统计学中相关系数性质的离散空间场相关指数模型来分析离散空间场之间的相关性,通过对CA预测的结果进行检验看,得出用离散空间场相关指数模型对CA模拟与预测的结果进行检验是一种有效的方法。更多还原

【Abstract】 Land use and Land Cover Change reflected the alteration of society and the ecological environment of an area. With the development of economy and acceleration of urbanization, its caused the land use and land Cover drastic change and holded the basic condition of the resources and environment all around in good time seems to be to very important. Fast dynamic monitoring to land use change made use of advanced technology of remote sensing and GIS, which can holded spatial information and time-series information about land use. Making use of the model of Cellular Automata based on the ideaistic of complex system to research the complex Behavior about the process of LUCC, discussing it quantitative and Forecasting the future condition can provide scientific basis for the area policymaker to valuate the condition of land use. Therefore, the typical region on the small spatial scale chosed to carry on remote sensing monitoring and space dynamic prediction of LUCC, which not only offered decision support for economic development in a region, but also riched the research of global land use change.Taking Jinan city as a case, configuration、process、mechanism and model study as core, the dissertation here carried out dynamic monitoring the land use and land cover change from 2000 to 2004 in research area by two TM images and the evolution tendency of land use in future was simulated and predicted, then analyzed the execution results by the computation of correlative index between discrete spatial fields based on entropy. The major contents and research results was as following: (1) It was a fast and efeicient method for monitoring to land use and land cover change by remote sensing image. In the study, the technology of RS and GIS was adapted to process the each image through preconditioning and enhancement handling, then adapted the method of extraction by supervised classification and visual judgement together, realized the higher precisional extraction of land use information. The results of the monitoring suggested that from 2000 to 2004 there was no significant change about land use and land cover except the urban land and the unused land in Jinan city;The total land use status transformed into the intensive land; In changed types, farmland transformed into other land due to advantageous geographic position and the difference of comparative economic benefit and its the main source of the increase of area of other land, the unused land distributed in beach area and it transformed into farm land mostly.(2) CA can simulate and express the distinct dynamic temporal-spatial variation of complex phenomena in the system through defining the relationship of cellular neighborhood and appropriate conversion rule and it can be applied in the research method for geographic system which contained the character of complex space-time. the dissertation here carried out simulating and predicting the land use change on 2008 and 2012 by the land-use and forecasting model. The results of the forecasting suggested that with time lapse, the area of farmland in Jinan city was reduced, the area of urban land was increasd and grew around the boundary, but there was no significant changes about forest land、grassland、water area and unused land.(3) the distributed data of land use and growth data of urban land simulated and predicted by CA were the fieid of discrete space which was continuation in uniform spaces and dispersion in attribute character, every kind of correlation index of space statistics and the degree of association index of grey system theory both can not calculate the correlation. The dissertation here constructed a computation of correlative index between discrete spatial fields based on entropy and calculate the correlation. The results of the forecasting by CA suggested that the computation of correlative index between discrete spatial fields was an effective method which checked the result simulated and predicted by CA.更多还原