大区域居住用地信息特征遥感影像提取方法研究

发布时间：2018-05-08 08:47

本文选题：居住用地 + 遥感影像　；参考：《中国地质大学（北京）》2013年博士论文

【摘要】：随着全球人口的增加和国民经济的发展,社会逐渐摆脱原有的以农业为主的情况,城镇化水平不断提高,越来越多的人涌入城市,城市随之增长和扩张。城市扩张往往牵涉到居住用地、农业用地和林地。农业用地和林地却常常无法抵挡城市扩张的汹汹势头。土地覆盖土地利用制图作为土地资源的基础数据,是世界各国各级政府、环保机构和私营企业所需要的。遥感影像由于其自身拥有如多分辨率,覆盖面广,可重复的观测和多／高光谱等技术优势,已经被广泛应用于从局部到全球尺度上的土地覆盖分类、目标识别和专题图制作,是一种可以进行大区域土地利用类型分类的经济、实用且可重复观测的方法。通过多时相分析,遥感能提供一个城市发展的独特视角。乡村和城市土地利用变化制图的一个关键因素是能区分乡村土地利用(农田、牧场、森林)和城市土地利用(住宅、商业、娱乐)。随着中国改革开放的逐步深化,经济建设的持续发展,城镇化水平的不断提高,居住用地面积日趋扩大。北京作为中国的首都,在20世纪末中国实行房地产市场改革以来,居住用地面积变化尤其明显。监控居住用地面积变化,首先必须获取客观准确的获得居住用地信息。本文以北京地区为研究区,基于Landsat卫星影像,分别采用非监督分类法、光谱分析法、监督分类法、面向对象的分类方法以及NDBI指数法提取出研究区的居住用地信息,然后对提取结果进行精度评价。基于监督分类的分类结果,结合实际调查数据,将研究区分为乡村居住用地和城镇居住用地。通过比较分析,发现不同的分类方法在乡村居住用地和城镇居住用地的分类效果不同。本研究结合多分类器并联组合的思想,将分类结果依据“胜者为王”的方法,将分类结果进行嵌套组合,最终得到比较准确的北京市居住用地信息。最后基于提取的北京市居住用地信息,对北京市三十年来居住用地时空分布的规模特征、居住用地时空分布的强度特征和居住用地时空分布的形态特征进行了简单分析。无论是五环内或者五环以外,北京市整体用地规模在不断扩大。五环内与五环内居住用地面积均在增加,居住用地斑块数整体在增加,出现减少的情况是因为若干小的斑块合并为较大的斑块。五环内的扩张强度大于五环外的扩展强度。本研究通过实验得出北京市五环内(市区)、五环外(村镇)居住用地用地规模扩大、用地扩展强度增加、形态特征上斑块面积增加,连片增大同时五环外2005年和2009年居住用地有零散增加趋势三个方面的变换特征,为进一步更好的模拟预测、优化提供有力支持。
[Abstract]:With the increase of the global population and the development of the national economy, the society gradually gets rid of the original situation of agriculture, the level of urbanization continues to improve, more and more people pour into the city, and the city grows and expands. Urban expansion often involves residential land, agricultural land and woodland. Agricultural land and woodland are often unable to resist the momentum of urban expansion. As the basic data of land resources, land cover mapping is needed by governments, environmental protection organizations and private enterprises all over the world. Remote sensing images have been widely used in land cover classification, target recognition and thematic mapping from local to global scale because of their own advantages such as multi-resolution, wide coverage, repeatable observation and multi-/ hyperspectral techniques. It is an economical, practical and repeatable method for classification of land use types in large areas. Through multi-temporal analysis, remote sensing can provide a unique perspective for urban development. A key factor in mapping rural and urban land-use change is the ability to distinguish between rural land use (farmland, pastures, forests) and urban land use (residential, commercial, recreational). With the gradual deepening of China's reform and opening up, the sustainable development of economic construction, the continuous improvement of urbanization level, the living land area is expanding day by day. Beijing is the capital of China. Since the reform of the real estate market in China in the late 20th century, the change of residential land area is especially obvious. To monitor the change of residential land area, we must obtain objective and accurate residential land information. In this paper, based on the Landsat satellite images in Beijing region, unsupervised classification, spectral analysis, supervised classification, object-oriented classification and NDBI index are used to extract the information of residential land in the study area. Then the accuracy of the extraction results is evaluated. Based on the results of supervised classification and actual survey data, the study is divided into rural residential land and urban residential land. Through comparative analysis, it is found that different classification methods have different classification effects on rural residential land and urban residential land. Combined with the idea of parallel combination of multiple classifiers, the classification results are nested and combined according to the method of "winner as king", and the more accurate information of residential land in Beijing is obtained. Finally, based on the extracted information of residential land in Beijing, this paper analyzes the characteristics of the spatial and temporal distribution of residential land in Beijing in the past 30 years, the intensity characteristics of space-time distribution of residential land and the morphological characteristics of space-time distribution of residential land. No matter inside the five rings or outside the five rings, the overall scale of land use in Beijing is constantly expanding. The area of residential land in the five rings and five rings is increasing, and the number of patches in the residential land is increasing as a whole. The decrease is due to the combination of several small patches into larger patches. The expansion strength in the five rings is greater than that outside the five rings. In this study, it is concluded that the scale of residential land, the intensity of land expansion and the area of patch in the urban area and outside of the five rings of Beijing are increased. At the same time, there are three transformation characteristics of the scattered increasing trend of residential land outside the five rings in 2005 and 2009, which provide powerful support for further better simulation prediction and optimization.
【学位授予单位】：中国地质大学（北京）
【学位级别】：博士
【学位授予年份】：2013
【分类号】：P237

【参考文献】