基于1992-2013年夜间灯光影像的内蒙古城市扩张动态分析研究

发布时间：2018-03-23 02:04

本文选题：DMSP/OLS　切入点：夜间灯光　出处：《内蒙古师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：地理国情监测是中国重要的基本国情,城市扩张研究和城镇化进程是它的重要体现。全国范围内的城镇化步伐的加快,使得内蒙古自治区的城镇化也在不断的推进,随着“3S”集成技术在城市遥感中的应用,越来越多的遥感手段开始应用到城市的发展和扩张分析中。现阶段的遥感已经形成了一个从地面到空间,从地理信息的数据收集、处理到拓扑叠加的分析与应用,成为获取资源环境信息与城市基础地理信息的重要手段。美国军事气象卫星(Defense Meteorological Satellite Program,DMSP)搭载的传感器(Operational Line-Scan System,OLS)获取的夜间灯光影像就是运用遥感手段获得的对全球的夜间灯光变化情况进行监测结果的数据表达。夜间的城市建成区灯光具有高亮度的特点,通过灯光灰度值(DN值)从0-63的变化可以进行建成区面积的提取,在面积提取的同时可以过滤掉乡村等干扰像元的干扰,对于城市的扩张分析研究起到事半功倍的作用。本研究通过获取1992-2013年22年的全球的DMSP/OLS夜间灯光影像数据,利用有理函数模型对夜间灯光影像数据进行年际校正和年间校正;以内蒙古自治区作为研究区,创新性的提出了最优阈值的选取方法:利用动态二分阈值法结合支持向量机(Support Vector Machine,SVM)分类算法进行历年建成区面积的提取;并进行空间叠加分析,绘制1992-2013年间的年代系列叠加图以及典型经济区城市市区建成区重心迁移角度和迁移距离图;最后分析影响城市扩张的驱动因子,得到如下结论:(1)通过概率统计和有理函数模型拟合计算校正参数,验证了《基于稳定映射点夜间灯光影像间相互辐射校正方法》对DMSP/OLS夜间灯光影像数据的高精度、快速和高效的校正。(2)针对蒙古高原地区夜间灯光亮暗的特点,创新出应用动态二分阈值法与支持向量机(SVM)分类算法相结合的方法来确定最优阈值;这两种方法结合使用能够快速、高精度的提取出内蒙古各个城市的建成区面积图斑。(3)通过对1992-2013年间城市的建成区形态变化进行叠加分析,结合城市自然条件和政策因素对典型经济区市区建成区扩张进行评价,结果表明:内蒙古自治区12个盟市的建成区随着近年来政策的扶持和经济的发展都出现了不同程度的扩张趋势;其中呼和浩特市市区在1996年后主要扩张方向为东南26.35°方向,1996-2008年重心的直接迁移距离22941m;包头市在2008年以后主要向东南48.47°方向发展扩张,重心直接迁移距离13888m;鄂尔多斯市呈“两城型”的模式扩张,2004-2013年市区建成区重心向西南方向105.53°的方向迁移了26788m。(4)由于内蒙古自治区盟市间扩张的不同步,通过对4个影响因子的相关分析和线性回归分析,寻找影响内蒙古自治区全区盟市扩张的主要驱动因子。其中单个影响因子做出最大贡献的是电力消费量;通过组合多元线性回归分析,总结出总人口数作为增强因子与电力消费量组合成为所有组合中对夜间灯光明暗变化影响最大的驱动因子。同时,以政府为主导的政策因子在城市的发展、扩张中起到了很大程度的驱动作用。
[Abstract]:Geographic conditions monitoring is an important basic national conditions Chinese, city expansion research and the process of urbanization is an important manifestation of it. The pace of urbanization nationwide accelerated, making the Inner Mongolia Autonomous Region urbanization is constantly advancing, with the application of "3S" integrated technology in remote sensing in the city, began the development of remote sensing means more and more with the expansion of city. The analysis is applied to the remote sensing stage has formed a from the ground to space, collected from geographic information data processing, analysis and application of topological overlay, has become an important means of access to information and resources and environment of city basic geographic information. The United States military meteorological satellite (Defense Meteorological Satellite Program. DMSP (Operational Line-Scan) equipped with sensors System, OLS) to obtain the nighttime light image is obtained using remote sensing technology to the global light at night The expression changes of the results of the monitoring data. The characteristics of night lighting city with high brightness, with light gray value (DN value) from 0-63 changes can be built extraction area, can filter off interference interference pixels at the same time the rural area extraction, for the expansion of the city to study multiplier effect. This study gets 1992-2013 years through 22 years of global DMSP/OLS nighttime light image data on the nighttime light image data correction and correction between interannual using rational function model; taking the Inner Mongolia Autonomous Region as the study area, the paper proposed a method to select the optimal threshold by two points: dynamic threshold method combined with support vector machine (Support Vector Machine, SVM) classification algorithm built over the extraction area; and spatial overlay analysis, drawing 1992-2013 years years Generation series superposition map and typical economic zone city urban migration angle and migration area centroid distance map; finally analyzes the driving factors affecting the city's expansion, we get the following conclusions: (1) through probability statistics and rational function model fitting calculation of correction parameters is verified based on the "stable mapping between the image point night light radiation correction method > on the DMSP/OLS nighttime light image data with high precision, fast and efficient correction. (2) according to the Mongolia plateau in the night lights bright and dark features, innovation and application of dynamic threshold method with two points of support vector machine (SVM) method combining classification algorithm to determine the optimal threshold; the combined use of these two methods fast, extract the built-up area of Inner Mongolia city of various patches with high precision. (3) overlay analysis by changing the built-up area form on the 1992-2013 years of the city, with the city natural conditions. A typical economic and policy factors of urban built-up area expansion area was evaluated, the results show that the the Inner Mongolia Autonomous Region 12 Union City built area in recent years with the development of policy and economy appeared expansion trend in different degree; the urban area of Hohhot city in 1996 after the main expansion direction is Southeast 26.35 degrees, 1996-2008 years of focus the migration distance of 22941m; the city of Baotou in 2008 after the main to the Southeast 48.47 degrees direction expansion, focus direct migration distance of 13888m; Erdos was "two city expansion mode" for 2004-2013 years, the urban built-up areas of gravity transfer 26788m. to the southwest direction 105.53 DEG (4) due to non synchronous the Inner Mongolia Autonomous Region League City the expansion, through the analysis of the 4 impact factor correlation analysis and linear regression, for the expansion of League City, the Inner Mongolia Autonomous Region region of the main drive for Son. The single influencing factor to make the greatest contribution to the power consumption; through the combination of multiple linear regression analysis, summed up the total population as the enhancement factor and the power consumption combination become the biggest driving factor of night light changes in all combinations. At the same time, the development of the policy factor is dominated by the government in the city the expansion, plays a driving role to a large extent.

【学位授予单位】：内蒙古师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P237

【参考文献】