晋陕蒙接壤区成长型资源城市土地利用变化遥感监测

发布时间：2018-09-09 09:43

【摘要】：晋陕蒙三省接壤区是我国煤炭基地,也是生态脆弱区。选择晋陕蒙接壤的山西省朔州市、陕西省榆林市及内蒙古鄂尔多斯市的主城区作为研究区,采用1995年、2002年、2007年以及2014年4个时相的Landsat TM/OLI数据,依据土地利用分类系统,采用随机森林分类方法对该城市不同时期的土地利用信息进行提取和分析。结果表明:(1)三座城市在近20年的城市发展过程中,土地利用变化剧烈。这种变化主要体现为朔城区建设用地比例从1.57%上升到11.71%;农用地的占地比例则由73.42%下降至69.71%;林地由22.94%逐年下降至15.61%;未利用地的变化从0.90%到1.92%。东胜区建设用地比例从0.53%上升到5.50%;农用地的占地比例由31.28%下降到21.85%;水体的占地比例由2.97%、上升至5.15%;未利用地由62.63%上升至65.33%。林地的面积变化不大。榆阳区建设用地占比从0.26%上升到1.82%;农用地的占地比例由13.98%上升到21.26%;水体由0.95%上升到3.86%;未利用地为82.41%降至70.04%。林地的面积在期间呈起伏变化,四年的占比分别为2.40%、5.23%、2.31%及3.02%。(2)本文采用随机森林方法进行土地利用分类,分类前采用多种软件对影像数据进行了处理,并波段运算了多种遥感指数作为特征变量来辅助分类。通过分类模型对四个时相三个研究区的遥感影像进行了土地利用分类,经过精度验证,采用500棵决策树保证了袋外误差均在0.2以下,kappa系数均在0.8以上,分类精度达到要求。而且整个分类过程的设计明晰,操作方便。但分类模型仍存在一些问题有待克服。(3)分析了三座城市的土地利用变化过程,分析了人口经济变化对土地利用的影响,并联系到它们作为成长型资源城市的定位和晋陕蒙煤炭产区的区位优势。本文认为当地土地利用变化的主要驱动力仍是第二产业尤其是煤炭产业,煤炭产业是当地的支柱产业,带动了当地的经济发展,并直接引起了当地土地利用结构的变化,以及所表现出的变化特征。
[Abstract]:Shanxi-Shaanxi-Mongolia three provinces border area is our country coal base, is also the ecological fragile area. The main urban areas of Shuozhou City, Shanxi Province, Yulin City, Shaanxi Province and Ordos City, Inner Mongolia were selected as the study areas. The Landsat TM/OLI data of 1995, 2002, 2007 and 2014 were used, and the land use classification system was used. The random forest classification method was used to extract and analyze the land use information in different periods of the city. The results show that: (1) during the development of the three cities in the past 20 years, the land use has changed dramatically. This change is mainly reflected in the increase of the proportion of construction land from 1.57% to 11.71%, the decrease of the proportion of land occupied by agricultural land from 73.42% to 69.71%, the decrease of forest land from 22.94% to 15.61%, and the change of unused land from 0.90% to 1.92%. The proportion of construction land in Dongsheng District increased from 0.53% to 5.50%; the proportion of land occupied by agricultural land decreased from 31.28% to 21.85%; the proportion of land occupied by water increased from 2.97% to 5.15%; and the proportion of unused land increased from 62.63% to 65.33%. There is little change in the area of woodland. The proportion of construction land in Yuyang District increased from 0.26% to 1.82%, the proportion of agricultural land occupied increased from 13.98% to 21.26%, the water increased from 0.95% to 3.86%, and the unused land decreased from 82.41% to 70.04%. The area of woodland fluctuates during the period, and the ratio of four years is 2.400.233.31% and 3.02% respectively. (2) in this paper, the land use classification is carried out by means of random forest method, and the image data are processed by many kinds of software before classification. In addition, several remote sensing indices are used as feature variables to assist classification. The land use classification of remote sensing images in four time phases and three study areas was carried out by the classification model. After accuracy verification, 500 decision trees were used to ensure that the out-of-bag errors were all below 0.2 and the kappa coefficients were all above 0.8, and the classification accuracy met the requirements. And the design of the whole classification process is clear and easy to operate. However, there are still some problems to be overcome in the classification model. (3) the process of land use change in three cities is analyzed, and the impact of population and economy change on land use is analyzed. And the location of them as a growing resource city and the regional advantages of Shanxi, Shaanxi and Mongolia coal production areas. This paper holds that the main driving force of local land use change is the secondary industry, especially the coal industry. The coal industry is the pillar industry of the local area, which drives the local economic development and directly causes the change of the local land use structure. And the characteristics of the changes.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P237

【相似文献】