【作者】 吴宏安；

【导师】 蒋建军；

【作者基本信息】 南京师范大学 ， 地图学与地理信息系统， 2006， 硕士

【摘要】 近年来，土地利用变化及其生态环境影响引起了日益广泛的重视，但是对于由城镇扩展所引起的生态问题研究还较少。本文以西安地区为研究区，基于1988年和2003年两期Landsat TM影像，提出了一种新的城镇信息提取方法，即比值居民地指数(Ratio Resident-area Index，RRI)法，提取了15年来该地区的城镇扩展信息。与最大似然监督分类法和人工神经网络分类法对比分析表明，RRI法能很好地去除裸地对城镇信息提取的影响，其精度也更高。此外，利用最大似然法获取了两期的土地利用图，通过土地利用面积转移矩阵分析了城镇用地与其它土地利用类型转化的关系，结合社会经济统计资料和当地环境部门的环境检测资料探讨了城镇扩展的生态环境影响。 通过本研究，得到了以下主要结论： (1) 利用比值居民地指数(RRI)，提取了西安地区1988年和2003年两期的城镇信息，并以此获得了该地区的城镇扩展信息。RRI法城镇提取精度比最大似然监督分类法和神经网络分类法的精度都要高，三者的总体精度分布为：87.08％，75.83％和79.17％。另外，15年间西安地区的城镇发展比较迅速，1988年的城镇用地面积(指居民地面积，非建成区面积)为87.251Km~2，而到2003年其面积则达132.482Km~2，增幅为51.84％。 (2) 通过面积转移矩阵，获取了城镇用地与其他土地利用类型的相互转化关系。分析表明，城镇的新增面积主要是由耕地、未利用地、果园和水体转化而来。其中耕地是城镇用地扩展的“主力军”。 (3) 西安地区城镇扩展的生态环境影响不容乐观。水资源“先天不足”和“后天污染”造成了该地区资源型和水质性“双重”缺水；随着城镇面积不断扩展，市区的“热岛效应”的倾向更加严重；城市需水量的增加，迫使城市超采地下水，使得西安地区的地面沉降和地裂缝等地质灾害表现的尤为突出。更多还原

【Abstract】 In recent years, the research on Land Use/Cover Change (LUCC) and its effects on eco-environment has attracted more and more scientists to pay attention to it. However, the ecological effects induced by urban expansion were seldom covered. The objective of this paper is to retrieve the information of urban expansion (1988-2003) in Xi’an region and its effects on eco-environment by using Landsat TM satellite images.In this study, two TM images of the study area (1988 and 2003) were used. After imagery pre-processing, which included atmospheric correction, geometric correction, and orthographic correction, ratio residential-area index (RRI) was proposed to retrieve the urban information. RRI is an index like RVI (ratio vegetation index), which can reflect the information of residential area. Compared with maximum likelihood classification (MLC) and artificial neural network (ANN) classification, the retrieval method of RRI could eliminate the influence of barrens, thus the accuracy of RRI was also higher than that of MLC and ANN. By analyzing the urban area in 1988 and 2003, the urban expansion of Xi’an region during the 15 years was obtained. In addition, supervised classification was used to retrieve land use information. The training samples were selected according to spectrum features, and the TM images of Xi’an region in 1988 and 2003 were classified into 6 land use types respectively, namely woodland, farmland, orchard, urban areas, water-body and uncultivated. With the help of area transfer matrix of land use changes, the relationship between urban expansion and the other five land uses was also analyzed. This could tell us where the expansion came from. At last, using related statistics data and environment detection data that we collected, the effects of urban expansion on eco-environment were analyzed.Through this study, the conclusions were drew as follows:(1) By using RRI, the information of urban expansion in Xi’an from 1988 to 2003 was acquired. And the overall accuracy of RRI, MLC and ANN were 87.08%, 75.83% and 79.17%, respectively. During the 15 years, the urban area of Xi’an region increased 51.84% from 87.251 Km~2 in 1988 to 132.482 Km~2 in 2003.(2) The urban expansion area was mainly transferred from farmland, uncultivated land, orchard and water-body. And farmland did the most important contribution to urban expansion.(3) The eco-environmental effects of urban expansion in the study area were not very optimistic. The pollution of water resources, Heat Island Effect and geological hazards were the most 3 obvious problems. Because of urban expansion and economic development, water pollution had induced water supply shortage in Xi’an region; With the continual increase of urban area which changed the ground feature, the annual average temperature of Xi’an city has increased 1.8℃ from 1989 to 2001; Urban expansion also needed more water supplies, which resulted in overexploiting ground water that caused many geological problems like ground subsidence and split.更多还原