基于TM图像和夜间灯光数据的区域城镇扩张监测—

引用本文

陈征, 胡德勇, 曾文华, 邓磊. 基于TM图像和夜间灯光数据的区域城镇扩张监测——以浙江省为例[J]. 国土资源遥感, 2014,26(1): 83-89
CHEN Zheng, HU Deyong, ZENG Wenhua, DENG Lei. TM image and nighttime light data to monitoring regional urban expansion: A case study of Zhejiang Province[J]. REMOTE SENSING FOR LAND & RESOURCES,2014,26(1): 83-89 复制到剪切板

Permissions

《国土资源遥感》编辑部

基于TM图像和夜间灯光数据的区域城镇扩张监测——以浙江省为例

陈征¹, 胡德勇¹, 曾文华², 邓磊¹

1.首都师范大学资源环境与旅游学院,北京 100048

2. 浙江省测绘资料档案馆,杭州 310012

胡德勇(1974-),男,副教授,主要从事遥感和地理信息系统在自然灾害、资源环境等领域的应用。Email:deyonghu@163.com。

第一作者简介: 陈征(1988-),女,硕士研究生,主要从事资源环境遥感研究。Email:lubyn@126.com。

收稿日期: 2013-02-28

基金: 国家973计划项目“LUCC对陆地生态系统的影响机制与多尺度LUCC生态效应研究”(编号: 2010CB950902)和国家自然科学基金项目(编号: 40901227)共同资助

摘要

以浙江省为研究区,利用收集到的Landsat TM图像、DMSP/OLS夜间灯光数据及其他辅助数据,开展区域城镇扩张监测方法的研究,并在此基础上分析了浙江省1995—2010年间城镇扩张的时空特征。结果表明: 基于分类回归树算法,综合利用TM数据和DMSP/OLS夜间灯光数据,可精确稳定地提取城镇信息; 1995—2010年间,浙江省城镇用地持续增加,除洞头县、庆元县、文成县、云和县、兰溪市、龙泉市和绍兴市等7个县市扩张加速度有所减缓外,全省其余市、县均处于加速扩张阶段; 浙江省城镇扩张地域差异明显,表现为沿海地区的城市化水平整体高于内陆地区,地势平坦地区高于地形复杂地区,从而形成了杭州、绍兴、温州3大沿海城市群和以金华为中心的内陆地区中小城市带。

关键词: Landsat TM; 夜间灯光数据; 城市扩张; 分类回归树算法; 浙江省

中图分类号:TP79 文献标志码:A 文章编号:1001-070X(2014)01-0083-07 doi: 10.6046/gtzyyg.2014.01.15

TM image and nighttime light data to monitoring regional urban expansion: A case study of Zhejiang Province

CHEN Zheng¹, HU Deyong¹, ZENG Wenhua², DENG Lei¹

1. College of Resource Environment and Tourism, Capital Normal University, Beijing 100048, China

2. Zhejiang Surveying and Mapping Data Archives, Hangzhou 310012, China

Abstract

Acquiring urban information and dynamically monitoring urban expansion forms constitute important parts of remote sensing technique in the field of resource and environment application. As a coastal province, Zhejiang has experienced a rapid economic development in the past 20 years. At the same time, its urban expansion phenomenon is significant. With Zhejiang Province as the study area, the authors obtained the spatial distribution information of urban land from Landsat TM data by using the CART algorithm. Under the condition of getting accurate classification results, the urban expansion spatial-temporal features of Zhejiang Province from 1995 to 2010 were analyzed.Applying CART to extracting urban information from TM data and optimized with NTL is an effective and adaptable method for monitoring regional urban expansion.During the past 15 years, the expansion acceleration of most cities is greater than zero, except for Dongtou, Qingyuan, Wencheng, Yunhe, Lanxi, Longquan and Shaoxing. From 1995 to 2010, Xiaoshan, Yuhang, Ningbo remained the first three cities in this aspect.It is shown that the urbanization levels of various cities are significantly different from each other. The urbanization level of coastal cities and cities in the terrain flat area is higher than that of the non-coastal cities and cities in the complicated topography area. As a result, the first three large coastal cities, which are Hangzhou, Shaoxing and Wenzhou, and the zone of small and medium sized cities in the non-costal area around Jinhua have been developed.

Keyword: Landsat TM; nighttime light data; urban expansion; classification and regression tree; Zhejiang Province

Show Figures

0 引言

近年来, 经济快速发展、人口向城镇聚集, 使得我国城镇用地面积不断增加, 范围不断扩大, 城市周边土地覆被类型不断变化, 大片乡村地区转化为城镇地区^[1]。因此对区域尺度上的城市扩张进行有效的监测可为土地规划和生态环境保护等提供科学参考, 具有重要的现实意义。

遥感技术具有多尺度、时效性强等特点, 已广泛应用于区域资源和环境监测领域。目前, 用于城镇用地信息提取的主要方法有图像分类法、阈值分割法等^{[2, 3, 4]}。然而, 由于受到遥感图像“ 同物异谱、同谱异物” 特征的影响, 阈值分割法在精确获取城镇专题信息过程中, 受像元值的一致性干扰较大, 需要对提取结果做进一步处理。20世纪90年代, DMSP/OLS夜间灯光数据(night-time light, NTL)作为一种新的数据源, 为动态监测城镇分布及其变化提供了新的方法和途径。运用NTL数据进行城镇扩张动态监测主要有灯光指数法^[5]和多层次阈值法^{[6, 7, 8]}等。前者综合考虑区域平均灯光强度和灯光面积构建灯光指数表征城市化水平, 利用阈值技术获取城镇空间分布特征及其变化; 后者通过构建灯光指数数据的多层次阈值, 对夜间灯光图像进行阈值分割, 获取城镇专题信息。但NTL数据空间分辨率较低, 对于城镇扩张监测存在局限性, 因而它常用作城镇扩张遥感信息提取的辅助数据源。以上方法虽然能够满足城镇变化监测的基本要求, 但如何应用多源数据提高城镇信息提取精度, 正确反映城镇扩张特征是目前亟待解决的问题。

本文以浙江省为研究区, 收集了1995年、2005年、2010(或部分2009)年 3个时期的TM图像, 同时收集了1995年、2005年、2010年的NTL数据和其他辅助数据, 在图像监督分类样本制备的基础上, 运用Breiman等^[9]提出的分类回归树(classification and regression tree, CART)算法进行图像分类, 提取了研究区各时期城镇用地空间分布专题信息; 为进一步提高专题信息提取的精度, 将NTL数据应用到城镇用地结果优化处理中, 获得了最终所需的城镇专题信息, 在此基础上采用GIS手段和统计方法宏观把握浙江省1995— 2010年城镇扩张的时空特征。

1 研究区概况及数据源

近20 a间, 浙江省经济快速发展, 土地利用类型变化明显, 变化最快的是城市用地和交通用地, 变化率分别达187.50%和110.03%^[10]。人口向城镇聚集, 大片乡村地区转化为城镇地区, 尤以杭州、宁波、浙中地区最为明显, 形成了以杭州为核心的大都市、以浙中为主要区域的中小城镇群发展模式。

为了对浙江省1995年后城市扩张情况进行监测, 本文选择了前述3个时期的TM图像, 空间分辨率为30 m。每时期各9景图像。数据覆盖范围如图1所示, 时相及特性如表1所示。

	Figure Option View Download New Window
	图1 研究区及Landsat TM数据景幅覆盖范围Fig.1 Study area and the coverage of Landsat TM data

表1 研究用到的Landsat TM数据及其特性 Tab.1 Landsat TM data and their characteristics

收集的NTL数据来源于NOAA下属的国家地球物理数据中心(national oceanic and atmospheric administration, NGDC, http: //www.ngdc.noaa.gov/dmsp/downloadV4composites.html)。NTL数据是美国国防气象卫星(defense meteorological satellite program, DMSP)搭载的线性扫描业务系统(operational line scan system, OLS)获取的, 该产品是一段时期内无云条件下灯光被连续探测到的频率, 并去除了亮云, 水体及火光等的影响^[11], 通过明暗对比反映城镇信息, 灰度范围0~63, 空间分辨率1 km。从NTL数据产品中, 提取稳定灯光数据, 经投影转换、裁剪等预处理后, 作为该研究区的城镇专题信息提取的辅助数据。

2 研究方法

2.1 城镇专题信息的遥感提取

城镇专题信息的遥感提取主要包括以下3个步骤:

1)图像分类数据准备。为了更好地区分植被、裸地和城镇用地, 以归一化差值植被指数(normalized difference vegetation index, NDVI)和归一化燃烧率(normalized burn ratio, NBR)作为遥感图像分类的辅助数据^{[12, 13]}, 即

NDVI=(NIR-R)/(NIR+R), (1)

NBR=(NIR-IR)/(NIR+IR), (2)

式中: R为TM第3波段(红光波段)亮度值; NIR为TM第4波段(近红外波段)亮度值; IR 为TM第7波段(短波红外波段)亮度值。

NDVI通过比值算法增强植被、耕地与其他地物的差别, 能够更好地反映植被的分布状况; NBR突出了植被变化前后的情况, 可以提高对于因植被数量变化而造成的裸露地表的提取精度。将TM图像的7个波段(热红外第6波段重采样为30 m)及其衍生数据(NDVI, NBR)组合为9个图层, 作为图像分类输入变量。

2)图像分类算法。研究选择CART作为图像分类的具体算法。该方法自1984年被提出以来, 在美国国家土地覆盖数据库(national land cover database, NLCD)开发中得到了推广和应用, 如NLCD 2000采用CART方法, 基于多种辅助数据层叠加分析, 依据各层的权重进行分类。CART算法是空间数据挖掘领域的一种常用分析方法, 其算法是一个二叉树递归过程, 利用训练样本建立基于规则的模型, 每一个规则集均定义了多元线性回归模型建立的条件。同时, 回归树模型能够说明目标变量和预测变量之间的非线性关系并允许连续变量和离散变量作为输入变量, 其分类精度优于其他简单线性回归模型 ^{[9, 14, 15, 16]}。

CART模型的精度取决于训练样本的质量。在实际操作中, 多景图像拼接后的镶嵌图像给训练样本的采集带来较大困难。为规避处理过程中的干扰因素, 本文采用分幅处理的方式提取单景TM图像中的城镇信息。此外, 区别于其他数据挖掘算法, CART模型需要采集较多的训练样本, 本文在每景图像选取50 000个像元以上城镇用地样本为测试变量, 同时保证交叉验证样本在5 000个像元以上。继而基于训练样本构建CART模型, 并利用该模型对待分类图层变量进行图像分类。

通常情况下, 乡村城市化转化是一个不可逆的过程, 多时序城镇专题信息提取及其城市扩张动态监测过程中, 前期城镇专题信息可以作为下一时期的训练样本, 从而保证获得足够的训练样本数量。

3)后处理建模及分类结果优化。基于CART算法输出的图像分类结果, 不可避免地存在错分或误分情况, 在研究区, 沿海滩地、裸地等对城镇用地信息提取的干扰尤为明显。为了减小乃至消除干扰因素对分类结果的影响, 进一步提高图像分类精度, 采用辅助数据进一步优化CART图像分类输出结果, 是后处理的常用手段。

针对研究区地表覆盖特性, 利用NTL数据对CART初步分类结果进行后处理和优化。其主要技术思路是选取最佳阈值范围内NTL数据与CART图像分类结果的叠置像元(公共像元), 并将其定义为实际的城镇专题信息。采用经验阈值与统计相结合的方法来确定NTL数据的城镇类别最佳阈值范围。综合TM图像、CART分类结果(像元值为DN_cart)和NTL(像元值为DN_ntl)3者叠加的统计结果, 结合Milesi^[17]的经验阈值设定动态阈值, 将CART 分类结果和NTL数据分别划分为3个层次(表2)。表中, 2 ≤ DN_cart ≤ 100, 0≤ DN_ntl≤ 63; m, n, j分别为2~100之间的1个数; l, k, f分别为1~63之间的1个数。以上6个参数根据实际情况取值, A, B, C为划分的3个层次。若提取满足A, B, C这 3个条件并与TM图像中城镇用地叠加误差最小, 则上述条件为最佳优化阈值范围。

表2 优化过程条件 Tab.2 Condition in the optimization process

2.2 技术流程

城市扩张动态监测的总体技术流程如图2所示。首先, 数据预处理包括去云处理、图像配准以及简单大气校正, 其中图像配准精度平坦地区控制在0.5个像元以内, 山区控制在1.5个像元以内。然后, 利用CART算法提取遥感图像的城镇专题信息, 主要包括训练样本制作、决策树构建和图像分类3个步骤; 期间辅以土地利用现状图、TM图像等为参考, 选取训练样本; 同时为保证样本的数量和质量, 须对选自土地利用现状图的样本进一步筛选, 仅保留置信度较高(如核心城镇区域)的样本, 应用CART算法提取TM图像中城镇信息, 并利用NTL数据对分类结果进行优化, 最终得到能够满足精度要求的分类结果。最后, 在分类结果满足要求的基础上, 结合GIS和统计分析方法对城镇扩张的时空特征进行分析。

	Figure Option View Download New Window
	图2 总体技术流程Fig.2 Overall flow chart

3 结果及精度验证

3.1 城镇信息提取结果

按照图2中的技术流程, 以单景TM图像为单位, 获取每景数据的CART算法运行结果, 然后采用NTL数据对该结果进行优化处理。在结果优化的过程中, 在每景图像的分类结果上采集随机点, 并对这些点所对应的分类结果像元值和灯光数据像元进行统计, 选取适当的阈值范围, 并进行结果优化。

在结果后处理过程中, 为消除细碎图斑对城镇制图的影响, 先滤除了少于5个像元的斑块, 得到单景遥感图像的专题结果; 然后对每个时相的9景数据进行镶嵌处理, 最终得到3个时相的城镇分布专题图, 如图3所示。

	Figure Option View Download New Window
	图3 1995— 2010年城镇用地空间分布示意图Fig.3 Spatial distribution of urban land from 1995 to 2010

3.2 精度检验

为了分析对比CART算法的图像分类精度和经过NTL数据优化后的专题信息提取精度, 进一步验证CART算法在图像分类中的优势, 本研究分别对2个过程的城镇信息提取结果进行了精度评估。在每景图像随机采集300个样本, 计算其信息提取的总体精度, 结果如表3所示。

表3 精度评价结果 Tab.3 Accuracy assessment results (%)

从表3可以看出, CART算法的图像总体分类精度达到80%以上, 且局部达到90%以上, 个别景受城镇用地分布不均的影响(如P120/R39景多数地段为山地, 城镇用地比重低)精度偏低。由此可见在训练样本采集的过程中, 样本选取对CART算法的运行精度存在一定的影响。

利用DMSP/OLS夜间灯光数据对CART运行输出结果进行优化后, 城镇专题信息提取的总体精度和优化前相比有大幅提升, 特别是CART算法运行输出结果精度较低的图像(P120/R39, P120/R40等)提升效果尤为明显, 所有图像的专题信息提取精度均达到85%以上, 平均值达到91.09%, 且数值波动较小, 表明提取结果具有一定的稳定性, 可以满足后续城镇扩张时空分析的数据要求。

4 城镇扩张分析

4.1 县级单位城镇扩张分析

城镇扩张特征可以从扩张面积(△ A)、扩张速度(U_v)和扩张强度(U_index)等方面分析, 表示为

△ A=U_i-U_j, (3)

U_v= $\begin{matrix} \frac{U_{i} - U_{j}}{i - j} \end{matrix}$ , (4)

U_index= $\begin{matrix} \frac{U_{v}}{U_{i}} \end{matrix}$ × 100%, (5)

式中: U_i, U_j为某一时间段起止城镇面积; i, j为起止年份。

扩张面积反映了某一时间段内城镇面积的增加总量, 扩张速度为城镇面积年均增长量^{[18, 19]}, 扩张强度是在某一段时间内年均城市面积增长比率^[20]。分别对研究区的城市扩张特征指数进行计算分析, 扩张速度较快的11个市的统计结果如表4所示。

表4 1995— 2005年、2005— 2010年各市扩张面积、扩张速度、扩张强度统计 Tab.4 Statistical analysis of expansion area, velocity and strength from 1995 to 2005 and from 2005 to 2010

通过对比图3可以发现: ①1995— 2005年浙江省城镇面积共增加1 992.1 km², 宁波市扩张速度最快, 其次为温州市和杭州市; ②2005— 2010年浙江省城镇面积共增加3 245.4 km², 宁波市和温州市继续保持高速扩张, 台州市扩张速度明显提高; ③总体说来, 1995— 2010年间, 浙江省城镇用地面积处于扩张态势, 除洞头县, 庆元县, 文成县, 云和县, 绍兴市, 兰溪市及龙泉市等7个县市扩张速度有所减缓外, 全省其余市、县均处于加速扩张阶段。

4.2 城市扩张空间特征

为定量描述浙江省城市扩张空间特征, 定义每个县级行政单元的城镇用地比率(R), 即

R= $\begin{matrix} \frac{A_{u}}{A_{t}} \end{matrix}$ × 100%, (6)

式中: R为城镇用地比率; A_u为某一行政单元城镇用地面积; A_t为该行政单元总面积。1995— 2010年城镇用地比率空间分布如图4所示。

	Figure Option View Download New Window
	图4 城镇用地比率空间分布图Fig.4 Spatial distribution of urban area ratio

从图4可以看出, 1995年研究区城市化水平总体较低, 最高城镇覆盖比率仅为29.2%; 2005年部分沿海城市城市化水平有明显提高, 其中绍兴市城镇覆盖比率超过50%, 杭州市提高到42.4%, 并且带动了周围城市的发展, 全省城市化水平较1995年有较大提高; 2010年城市化水平全面提高, 绍兴市、杭州市城镇覆盖比率达50%以上, 全省有20个县市城镇覆盖比率超过10%。

1995— 2010年, 沿海地区的城市化水平整体高于内陆地区, 地势平坦地区高于地形复杂地区, 形成了杭州, 绍兴, 温州3大沿海城市群和以金华为中心的内陆地区中小城市带。

5 结论

本文以浙江省为研究区, 利用CART算法对1995年、2005年及2010年3个时期的TM图像进行分类, 并利用DMSP/OLS夜间灯光数据对分类结果进行了优化, 在此基础上利用GIS和统计方法对1995— 2010年间浙江省城镇扩张特征进行统计和分析, 主要结论如下:

1)基于CART算法, 综合利用TM图像和DMSP/OLS夜间灯光数据, 是适用于区域尺度城镇扩张监测的一种有效且稳定性较好的信息提取方法。本研究中, 采用CART算法以TM图像作为单一数据源进行信息提取, 得到1995年、2005年及2010年总体分类精度分别为83.3%, 78.5%和81.1%, 但由于受地形和图像光谱特征的影响, 分类精度偏低且不够稳定; 增加DMSP/OLS夜间灯光数据对分类结果进行优化, 提高了分类精度及其稳定性, 平均精度分别达到92.0%, 91.1%和90.1%。

2)1995— 2010年间, 浙江省城镇用地持续增加, 浙江省城镇用地面积处于扩张态势。除洞头县, 庆元县, 文成县, 云和县, 兰溪市, 龙泉市及绍兴市等7个县市扩张速度有所减缓外, 全省其余市、县、区均处于加速扩张阶段。其中, 1995— 2005年城镇扩张面积达1 992.1 km², 2005— 2010年城镇扩张面积为3 245.4 km², 宁波市, 温州市, 杭州市和台州市均处于快速扩张阶段。

3)城市扩张的空间特征方面, 沿海地区的城市化水平整体高于内陆地区, 地势平坦地区高于地形复杂地区, 形成了杭州、绍兴、温州3大沿海城市群和以金华为中心的内陆地区中小城市带。

由于TM图像和NTL数据的空间分辨率存在较大差异(TM数据空间分辨率为30 m, 而NTL空间分辨率为1 km), 造成在结果优化的过程中小面积城镇信息的遗漏, 对城镇扩张的动态监测结果的准确性带来一定的影响, 如何克服这一问题也成为今后本研究的工作重点。

The authors have declared that no competing interests exist.

参考文献

View Option

[1]	胡德勇, 李京, 陈云浩, 等. 基于多时相Land sat数据的城市扩张及其驱动力分析[J]. 国土资源遥感, 2006, 18(4): 46-49, 54. Hu D Y, Li J, Chen Y H, et al. An analysis of urban expansion and its dynamics based on multi-temporal Land sat data[J]. Remote Sensing for Land and Resources, 2006, 18(4): 46-49, 54. [本文引用:1]
[2]	查勇, 倪绍祥, 杨山. 一种利用TM图像自动提取城镇用地信息的有效方法[J]. 遥感学报, 2003, 7(1): 37-40. Zha Y, Ni S X, Yang S. An effective approach to automatically extract urban land -use from TM imagery[J]. Journal of Remote Sensing, 2003, 7(1): 37-40. [本文引用:1] [CJCR: 0.992]
[3]	徐涵秋. 基于谱间特征和归一化指数分析的城市建筑用地信息提取[J]. 地理研究, 2005, 24(2): 311-319. Xu H Q. Fast information extraction of urban built-up land based on the analysis of spectral signature and normalized difference index[J]. Geographical Research, 2005, 24(2): 311-319. [本文引用:1]
[4]	潘卫华, 徐涵秋. 泉州市城市扩展的遥感监测及其城市化核分析[J]. 国土资源遥感, 2004, 16(4): 36-40. Pan W H, Xu H Q. A study of urban spatial expansion of Quanzhou City on the basis of remote sensing technology and urbanization core analysis[J]. Remote Sensing for Land and Resources, 2004, 16(4): 36-40. [本文引用:1]
[5]	卓莉, 史培军, 陈晋, 等. 20世纪90年代中国城市时空变化特征——基于灯光指数CNLI方法的探讨[J]. 地理学报, 2003, 58(6): 893-901. Zhuo L, Shi P J, Chen J, et al. Application of compound night light index derived from DMSP/OLS data to urbanization analysis in China in the 1990s[J]. ACTA Geographica Sinica, 2003, 58(6): 893-901. [本文引用:1] [CJCR: 2.937]
[6]	徐梦洁, 陈黎, 刘焕金, 等. 基于DMSP/OLS夜间灯光数据的长江三角洲地区城市化格局与过程研究[J]. 国土资源遥感, 2011, 23(3): 106-112. Xu M J, Chen L, Liu H J, et al. Pattern and process of urbanization in the Yangtze Delta based on DMSP/OLS data[J]. Remote Sensing for Land and Resources, 2011, 23(3): 106-112. [本文引用:1]
[7]	Liu Z F, He C Y, Zhang Q F, et al. Extracting the dynamics of urban expansion in China using DMSP-OLS nighttime light data from 1992 to 2008[J]. Land scape and Urban Planning, 2012, 106(1): 62-72. [本文引用:1] [JCR: 2.314]
[8]	Zhang Q L, Seto K C. Mapping urbanization dynamics at regional and global scales using multi-temporal DMSP/OLS nighttime light data[J]. Remote Sensing of Environment, 2011, 115(9): 2320-2329. [本文引用:1] [JCR: 5.103]
[9]	Breiman L, Friedman J, Olshen R, et al. Classification and regression trees[M]. New York: Chapman and Hall, 1984: 358. [本文引用:2]
[10]	赵哲远, 马奇, 华元春, 等. 浙江省1996—2005年土地利用变化分析[J]. 中国土地科学, 2009, 23(11): 55-60, 54. Zhao Z Y, Ma Q, Hua Y C, et al. Analysis on land use changes from 1996 to 2005 in Zhejiang Province[J]. China Land Science, 2009, 23(11): 55-60, 54. [本文引用:1] [CJCR: 2.031]
[11]	王翠平, 王豪伟, 李春明, 等. 基于DMSP/OLS影像的我国主要城市群空间扩张特征分析[J]. 生态学报, 2012, 32(3): 942-954. Wang C P, Wang H W, Li C M, et al. Analysis of the spatial expansion characteristics of major urban agglomerations in China using DMSP/OLS images[J]. Acta Ecologica Sinica, 2012, 32(3): 942-954. [本文引用:1]
[12]	Alphan H, Derse M A. Change detection in southern Turkey using normalized difference vegetation index(NDVI)[J]. Journal of Environmental Engineering and Land scape Management, 2013, 21(1): 12-18. [本文引用:1] [JCR: 1.041]
[13]	Epting J, Verbyla D, Sorbel B. Evaluation of remotely sensed indices for assessing burn severity in interior Alaska using Land sat TM and ETM+[J]. Remote Sensing of Environment, 2005, 96(3-4): 328-339. [本文引用:1] [JCR: 5.103]
[14]	Yang L M, Huang C Q, Homer C G, et al. An approach for mapping large-area impervious surfaces: Synergistic use of Land sat-7 ETM+ and high spatial resolution imagery[J]. Canadian Journal of Remote Sensing, 2003, 29(2): 230-240. [本文引用:1] [JCR: 0.986]
[15]	刘勇洪, 牛铮, 王长耀. 基于MODIS数据的决策树分类方法研究与应用[J]. 遥感学报, 2005, 9(4): 405-411. Liu Y H, Niu Z, Wang C Y. Research and application of the decision tree classification using MODIS data[J]. Journal of Remote Sensing, 2005, 9(4): 405-411. [本文引用:1] [CJCR: 0.992]
[16]	Huang C, Townshend J R G. A stepwise regression tree for nonlinear approximation: Applications to estimating subpixel and cover[J]. International Journal of Remote Sensing, 2003, 24(1): 75-90. [本文引用:1]
[17]	舒松, 余柏蒗, 吴健平, 等. 基于夜间灯光数据的城市建成区提取方法评价与应用[J]. 遥感技术与应用, 2011, 26(2): 169-175. Shu S, Yu B L, Wu J P, et al. Methods for deriving urban built-up area using night-light data: Assessment and application[J]. Remote Sensing Technology and Application, 2011, 26(2): 169-175. [本文引用:1] [CJCR: 0.749]
[18]	Fan F L, Wang Y P, Qiu M H, et al. Evaluating the temporal and spatial urban expansion patterns of Guangzhou from 1979 to 2003 by remote sensing and GIS methods[J]. International Journal of Geographical Information Science, 2009, 23(11): 1371-1388. [本文引用:1] [JCR: 1.613]
[19]	林目轩, 师迎春, 陈秧分, 等. 长沙市区建设用地扩张的时空特征[J]. 地理研究, 2007, 26(2): 265-274. Lin M X, Shi Y C, Chen Y F, et al. A study on spatial-temporal features of construction land expansion in Changsha urban area[J]. Geographical Research, 2007, 26(2): 265-274. [本文引用:1]
[20]	穆晓东, 刘慧平, 薛晓娟. 基于遥感监测的北京1984—2007年城市扩展研究[J]. 北京师范大学学报: 自然科学版, 2012, 48(1): 81-85. Mu X D, Liu H P, Xue X J. Urban growth in Beijing from 1984 to 2007 as gauged by remote sensing[J]. Journal of Beijing Normal University: Natural Science, 2012, 48(1): 81-85. [本文引用:1] [CJCR: 0.4356]

2006

0.0

胡德勇, 李京, 陈云浩, 等. 基于多时相Land sat数据的城市扩张及其驱动力分析[J]. 国土资源遥感, 2006, 18(4): 46-49, 54.

Hu D

, Li

, Chen Y

, et al. An analysis of urban expansion and its dynamics based on multi-temporal Land sat data[J]. Remote Sensing for Land and Resources, 2006, 18(4): 46-49, 54.

Based on multi-temporal Landsat imageries, the authors extracted the urban built-up areas in Changsha City, Hunan province on the basis of such methods as supervised classification, unsupervised classification and NDBI index. The spatio-temporal patterns of the urban expansion were analyzed quantitatively by using the entropy value and differential entropy value of sequential urban land use images. Overlay and buffer spatial analysis were used to establish the relationship between the urban expansion and the distance from the city center. On such a basis, the expansion rate map was drawn to show the dominant direction of urban expansion. The results show that the urban built-up areas have continued to increase, and the expansion was directed mostly southeastward from 1973 to 1986, westward from 1987 to 1993, and southward and southeastward from 1994 to 2001. The rapid increase of population is thought to be the most important driving force of urban expansion.

以长沙市为例，在多时相Landsat遥感数据支持下，采用监督分类、非监督分类和归一化裸露指数(NDBI)等方法提取城市用地信息。通过对比多期城市用地的熵值变化，定量分析城市扩张的时空特性；运用叠加、缓冲区分析等方法，分析城市扩张和中心城区的关系，绘制城市扩张速度玫瑰图。研究结果表明，长沙市建成区总面积不断扩展，其中，1973~1986年扩张主要表现在东南方向，1987~1993年为西部方向，1994~2001年南和东南方向成为快速扩张方向；对长沙城市扩张驱动力进行分析，认为人口迅速增加是城市扩张的最主要驱动力。

... 0 引言近年来,经济快速发展、人口向城镇聚集,使得我国城镇用地面积不断增加,范围不断扩大,城市周边土地覆被类型不断变化,大片乡村地区转化为城镇地区^[1] ...

2003

0.0

0.992

查勇, 倪绍祥, 杨山. 一种利用TM图像自动提取城镇用地信息的有效方法[J]. 遥感学报, 2003, 7(1): 37-40.

Zha

, Ni S

, Yang

An effective approach to automatically extract urban land -use from TM imagery[J]. Journal of Remote Sensing, 2003, 7(1): 37-40.

Land covers in urban areas tend to change drastically over a short period of time due to rapid urbanization. Especially, land covers in China are changing more frequently, which have been occurred in urban and peri-urban areas where accompanied rapid economic development. Remote sensing images are ideally used to monitor current land cover changes thanks to their rapid up-date capability. There are two conventional methods in producing land covers from either aerial photographs or satellite images: manual interpretation and computer classification. The conventional methods of converting images into urban land covers have a number of limitations. The first method is tedious and highly subject to the interpretation of the image analyst, and fails to take advantage of all spectral information contained in multi-spectral remote sensing images because only limited number of spectral bands can be used to generate the false color composite. The second method, even though requires much less human intervention, is still subject to many limitations. It is difficult task to achieve a high accuracy level, and the conventional supervised classification method requires the selection of a number of training samples for all those covers whose representativeness directly affects the classification accuracy. In this study we present a new method based on Normalized Difference Built-up Index(NDBI): NDBI=(TM5-TM4)/(TM5+TM4) This method has been successfully applied to automatically extract the urban land in Wuxi City, Eastern China by manipulating the spectral bands of TM imagery. The results indicate that it can serve as a worthwhile alternative for quickly extracting urban land. This proposed new method has a number of advantages over the conventional supervised classification method. It is very objective because it requires minimun interference from the image analyst, much faster and may acquire a high accuracy. However, the universality of this proposed method needs to be tested further in other geographic areas as the spectral response of the various land covers may vary spatially.

如何快速、准确与客观地提取城镇用地信息，以获得城镇用地的分布范围和面积资料，是有关城镇问题研究中经常涉及到的一个基本问题，现代遥感技术为这个问题的有效解决提供了强有力的保证，运用提出的归一化建筑指数，从TM图像进行了无锡市城镇用地信息的自动提取，研究结果表明，与传统的计算机分类和手工屏幕数字化方法相比，归一化建筑指数法是一种非常行之有效的方法。

... 目前,用于城镇用地信息提取的主要方法有图像分类法、阈值分割法等^[2,3,4] ...

2005

0.0

徐涵秋. 基于谱间特征和归一化指数分析的城市建筑用地信息提取[J]. 地理研究, 2005, 24(2): 311-319.

Xu H

Fast information extraction of urban built-up land based on the analysis of spectral signature and normalized difference index[J]. Geographical Research, 2005, 24(2): 311-319.

This paper studies the principles and method of fast remote-sensing information extraction for urban built-up land, taking Fuzhou city as an example. With the detailed analysis and clarity of several existing normalized difference indices, the study selects three indices, Normalized Difference Built-up Index (NDBI), Modified Normalized Difference Water Index (MNDWI) and Soil Adjusted Vegetation Index (SAVI) to represent three major urban land use/cover classes, i. e. , built-up land, water body and vegetation, respectively. The three index images are generated from a Landsat ETM + subscene of Fuzhou city and then used as three bands to compose a new image. This dramatically compresses the original eight-band ETM+ image into a three-band image, reduces band correlations and data redundancy, thus significantly simplifiying the band spectral analysisprocedures. The spectral signature analysis only needs to be performed on this three-index composite image and the signature differences among the three major urban land use/cover classes are revealed much easier than being done with multi-bands. Based on the revealed signature differences, the built-up land is finally extracted through a simple logic calculation. The result achieves a 91.3% overall accuracy. Therefore, the method is a fast and accurate one for the remote-sensing information extraction of urban land use without human interference. In addition to the above built-up land information extraction study, the paper proposes a Modified Normalized Difference Water Index (MNDWI) based on the NDWI of Mcfeeters (1996) , which uses MIR wavelength (ETM + band 5) instead of NIR wavelength to construct the index. The replacement largely enhances the contrast of the water bodies with the other land use/cover classes and reduces the spectral confusion with the other classes. Therefore, the MNDWI is more suitable for delineating features of polluted urban rivers/lakes. The advantages of using SAVI instead of NDVI in the urban study are also discussed in this paper.

以福州市ETM+影像为例,研究了城市建筑用地信息快速准确提取的原理和方法。通过对归一化差异型指数构成原理的分析以及对同名异义和异名同义现象的甄别,选取了归一化差异建筑指数(NDBI)、修正归一化差异水体指数(MNDWI)和土壤调节植被指数(SAVI)来代表城市建成区的三种最主要的土地利用类型--建筑用地、水体和植被。在此基础上进一步对这三个新的指数波段进行谱间特征分析,最后利用基于规则的逻辑判别运算将城市建筑用地信息提取出来。研究表明这一方法可以使繁杂的多波段谱间分析得以简化, 是一种快速准确、未经人工干预的建筑用地信息提取方法。本文还探讨了在城市建成区的研究中采用SAVI指数替代NDVI指数的优点。

... 目前,用于城镇用地信息提取的主要方法有图像分类法、阈值分割法等^[2,3,4] ...

2004

0.0

潘卫华, 徐涵秋. 泉州市城市扩展的遥感监测及其城市化核分析[J]. 国土资源遥感, 2004, 16(4): 36-40.

Pan W

, Xu H

A study of urban spatial expansion of Quanzhou City on the basis of remote sensing technology and urbanization core analysis[J]. Remote Sensing for Land and Resources, 2004, 16(4): 36-40.

Urban construction and development constitute the major symbols of social progress, and the spatial data of urban expansion make up an important index in evaluating urbanization. To monitor the change of the urban area, the authors employed remote sensing technology and Landsat TM/ETM + images of 1989 and 2000 to extract the urban expansion information of Quanzhou City within the eleven years. The methods included the Normalized Difference Built-up Index (NDBI) and the image supervised classification. The spatial information was combined with the socio-economic data through GIS, which could be used in the analysis of urbanization core of Quanzhou City by means of the urbanization theory. The study reveals that the driving force of the urbanization process in Quanzhou City was the rapid development of village and township enterprises during the study period.

以泉州市为研究区,采用仿归一化植被指数(NDBI)和监督分类法对1989年和2000年的TM/ETM + 图像进行城市空间信息提取,分析了城市空间扩展的趋势,探索出泉州市是以乡镇企业为城市化核来进行城市化过程的

... 目前,用于城镇用地信息提取的主要方法有图像分类法、阈值分割法等^[2,3,4] ...

2003

0.0

2.937

卓莉, 史培军, 陈晋, 等. 20世纪90年代中国城市时空变化特征——基于灯光指数CNLI方法的探讨[J]. 地理学报, 2003, 58(6): 893-901.

Zhuo

, Shi P

, Chen

, et al. Application of compound night light index derived from DMSP/OLS data to urbanization analysis in China in the 1990s[J]. ACTA Geographica Sinica, 2003, 58(6): 893-901.

This paper presented compounded night light index (CNLI) derived from DMSP/OLS data which were pre-processed by Japan National Institute of Environmental Studies for regional urbanization level estimation considering the light spatial distribution and intensity. The correlation analyses between light indices and composite urbanization indices were carried out at provincial level and county level. They can also be carried out in other scales if appropriate census data are available. The results show that there were significant relationships between the two kinds of indices. The regression models at provincial scale and county scale for composite urbanization index estimation using CNLI were also developed. It suggests that CNLI is an effective and applicable index for regional urbanization analysis and monitoring. Analysis of the changes of CNLI, S and I in China during the period 1992-1998 showed that the imbalance of urbanization level in China was obvious. It was higher in the east and lower in the west. Urbanization level was improved largely during 1992-1998, but the speed and the type of development were different.

1. Key Laboratory of Environmental Change and Natural Disaster, Ministry of Education of China, Beijing Normal University, Beijing 100875, China; 2. Japan National Institute of Environmental Studies, 16-2 Onogawa, Tsukuba, 305-8506, Japan

用DMSP/OLS夜间平均灯光强度图像在区域上的扩展特征和聚集特征构建了一反映区域城市化综合水平的灯光指数CNLI (compounded night light index) , 并分别在省级和县级尺度上对CNLI与基于统计数据的城市化水平指数做了相关分析，结果表明CNLI与基于统计数据的城市化水平复合指数显著相关，最后用CNLI分析了中国省际城市化水平的时空特征。CNLI为客观、实时监测城市化水平的空间格局及其发展变化提供了新的指标和方法。

... 运用NTL数据进行城镇扩张动态监测主要有灯光指数法^[5]和多层次阈值法^[6,7,8]等 ...

2011

0.0

徐梦洁, 陈黎, 刘焕金, 等. 基于DMSP/OLS夜间灯光数据的长江三角洲地区城市化格局与过程研究[J]. 国土资源遥感, 2011, 23(3): 106-112.

Xu M

, Chen

, Liu H

, et al. Pattern and process of urbanization in the Yangtze Delta based on DMSP/OLS data[J]. Remote Sensing for Land and Resources, 2011, 23(3): 106-112.

The lighted area, night light intensity, compounded night light index (CNLI) and total night light index (TNLI) of 16 cities in the Yangtze Delta were derived from the DMSP/OLS night light data in 2008, and their correlations with urbanization statistics were calculated respectively. Of these factors, the night light intensity proved to be an effective indicator for assessing the urbanization level. Construction land information of urban agglomerations in 1998, 2003 and 2008 was also extracted from DMSP/OLS night light data after the determination of the threshold value. Then the pattern and process of urbanization in the Yangtze Delta were analyzed and the expansion process of buildup area was categorized into several types. During the study period, urban agglomerations in the Yangtze Delta had three typical spatial expansion modes, of which the polygon mode could be observed in all the cities, the linear mode exists along main highways and railways, and the point mode occurs mainly around the less important cities. In spite of the imbalance of urbanization level within the study area, the regional differentiation was lessened gradually. According to the variation of the threshold value from 1998 to 2008, 11 cities, namely Zhenjiang, Changzhou, Yangzhou, Wuxi, Suzhou, Nantong, Huzhou, Jiaxing, Shaoxing, Zhoushan and Taizhou, belonged to the filling-in dominant type, Nanjing, Hangzhou and Shanghai belonged to the first-extension-then-filling-in dominant type, and Taizhou and Ningbo belonged to the first-filling-in-then-extension dominant type.

利用1998、2003和2008年的DMSP/OLS夜间灯光数据,结合相关统计数据,分析灯光数据与城市化进程中经济、人口和土地利用等方面的关系,构建基于灯光数据的城市化水平指数,分析长三角城市群城市化进程和空间模式; 通过与统计数据对比,按照最小误差原则确定市辖区建设用地信息提取的阈值,根据各城市在1998~2003年和2003~2008年两个阶段建设用地信息提取阈值的变化,划分城市建设用地的主导扩张类型。研究表明: 长三角城市群表现为核心城市的面状扩张、沿交通网络的线状扩张和非核心城市的点状扩张模式,城市群的特征逐渐体现; 区域内部城市化发展水平不均衡,但空间差异逐渐缩小; 镇江、常州、扬州、无锡、苏州、南通、湖州、嘉兴、绍兴、舟山和台州等市属于填充增强主导扩展类型; 南京、杭州和上海则属于外延-填充增强主导扩展类型; 泰州和宁波属于填充增强-外延主导扩展类型。

... 运用NTL数据进行城镇扩张动态监测主要有灯光指数法^[5]和多层次阈值法^[6,7,8]等 ...

2012

2.314

0.0

... 运用NTL数据进行城镇扩张动态监测主要有灯光指数法^[5]和多层次阈值法^[6,7,8]等 ...

2011

5.103

0.0

... 运用NTL数据进行城镇扩张动态监测主要有灯光指数法^[5]和多层次阈值法^[6,7,8]等 ...

1984

0.0

... 本文以浙江省为研究区,收集了1995年、2005年、2010(或部分2009)年 3个时期的TM图像,同时收集了1995年、2005年、2010年的NTL数据和其他辅助数据,在图像监督分类样本制备的基础上,运用Breiman等^[9]提出的分类回归树(classification and regression tree,CART)算法进行图像分类,提取了研究区各时期城镇用地空间分布专题信息 ...

... 同时,回归树模型能够说明目标变量和预测变量之间的非线性关系并允许连续变量和离散变量作为输入变量,其分类精度优于其他简单线性回归模型 ^[9,14,15,16] ...

2009

0.0

2.031

... 03%^[10] ...

2012

0.0

王翠平, 王豪伟, 李春明, 等. 基于DMSP/OLS影像的我国主要城市群空间扩张特征分析[J]. 生态学报, 2012, 32(3): 942-954.

Wang C

, Wang H

, Li C

, et al. Analysis of the spatial expansion characteristics of major urban agglomerations in China using DMSP/OLS images[J]. Acta Ecologica Sinica, 2012, 32(3): 942-954.

Since the 1990s, the urban sprawl phenomenon has become very obvious in China due to the pace of urbanization. We exacted urban pixel of Jingjinji, Yangtze River Delta and Pearl River Delta urban agglomerations based on DMSP/OLS images in 1992, 1995, 2000, 2005 and 2009. We also analyzed the spatial expansion characteristics of urban land in the three agglomerations. We calculated the spatial expansion intensity and spatial expansion type in the three agglomerations. Then we analyzed the directional spatial expansion characteristics and compared them to the level of socio-economic development (population urbanization rate and the proportion of secondary and tertiary industry in GDP) in the cities within the three urban agglomerations. The results showed that (1) the intensity of spatial expansion in the Yangtze River Delta was higher than the other regions during the early and middle stages of spatial expansion, the intensity of spatial expansion in Jingjinji was higher than in the other regions during the late stage of urban expansion, and the intensity of urban expansion in the Pearl River Delta was relatively weak. (2) The ‘infilling’ type of spatial expansion was observed in all three urban agglomerations until 2009. (3) Urban land in Jingjinji expanded east, south and southeast from the centers of Beijing, Tianjin and Tangshan. Urban land in the Yangtze River Delta expanded west, southwest and northwest from the center of Shanghai, and expansion in local areas was centered around several urban patches, and urban land in the Pearl River Delta expanded around the central urban clusters. (4) From 1992 to 2009, the social economic development level, population urbanization rate and the proportion of secondary and tertiary industry in GDP in Beijing, Tianjin and Tangshan were higher than the other cities in the Jingjinji. The population urbanization rate and the proportion of secondary and tertiary industry in GDP of Shanghai, Nanjing, Wuxi and Suzhou were higher than the other cities in the Yangtze River Delta, and the population urbanization rate and the proportion of secondary and tertiary industry in GDP of Guangzhou, Shenzhen, Zhongshan, Zhuhai, Dongguan and Foshan were higher than the other cities in the Pearl River Delta. These results all indicated that rapid socioeconomic development drives the expansion of urban land. (5) The general spatial expansion trend of urban land in the three agglomerations was from dispersed to cluster form, and the urban land gradually expanded as a banded or clustered interlocking region. We conclude that urban land in China forms clusters as a result of rapid socioeconomic development, the characteristics of urban land expansion varies in different agglomerations, and the analysis provides data support for continued urbanization in China.

20世纪90年代以来,随着中国城市化步伐的加快,城市用地空间扩张极为明显。采用1992、1995、2000、2005、2009年5期的夜间灯光影像数据提取了京津冀、长江三角洲和珠江三角洲城市群的城市像元,并从空间扩张强度、扩张类型以及方向性空间扩张特征3个方面对三大城市群城市用地的空间扩张特征进行了分析,同时还从社会经济方面对城市群空间扩张的驱动因素进行了讨论。结果表明,三大城市群的空间扩张特征既存在共性也存在明显的差异,同时社会经济的快速发展对城市用地的扩张具有明显的驱动作用,为进一步推动中国城市化进程提供了数据支撑。

... NTL数据是美国国防气象卫星(defense meteorological satellite program,DMSP)搭载的线性扫描业务系统(operational line scan system,OLS)获取的,该产品是一段时期内无云条件下灯光被连续探测到的频率,并去除了亮云,水体及火光等的影响^[11],通过明暗对比反映城镇信息,灰度范围0~63,空间分辨率1 km ...

2013

1.041

0.0

... 为了更好地区分植被、裸地和城镇用地,以归一化差值植被指数(normalized difference vegetation index,NDVI)和归一化燃烧率(normalized burn ratio,NBR)作为遥感图像分类的辅助数据^[12,13],即 ...

2005

5.103

0.0

2003

0.986

0.0

2005

0.0

0.992

刘勇洪, 牛铮, 王长耀. 基于MODIS数据的决策树分类方法研究与应用[J]. 遥感学报, 2005, 9(4): 405-411.

Liu Y

, Niu

, Wang C

Research and application of the decision tree classification using MODIS data[J]. Journal of Remote Sensing, 2005, 9(4): 405-411.

Decision tree classification algorithms have significantpotential in remote sensing data classification. In this\nresearch, two popular decision tree algorithms―――CART and C4・5 are presented, and two techniques known as boosting\nand bagging inmachine learning area are introduced. We examined thesemethods tomaxmi ize classification accuracies\nusing these decision trees and techniques tomap land cover ofHuabei area in China fromMODIS 250m data. The result\nindicates that decision tree with abundance training samples has higher classification accuracy than maxmi um likelihood\nclassifier(MLC) in the land cover classification test, whereas insufficientsamples resulted in a lower accuracy fordecision\ntree than MLC. The result also shows CART algorithm has more advantageous than C4・5 algorithm in classification\naccuracy and tree structure. And the decision tree classification accuracy depends on the optmi al structure and pruning\nprocess. We also tested the behaviour ofboosting and bagging techniques combinedwith CART and the result shows that\nadding boosting technique to decision tree can increase classification accuracies by 18・5%―25・6% for the poorly\nseparable classes inMLC.

介绍了目前国际上流行的两种决策树算法―――CART算法与C4?5算法,并引入了两种机器学习领域里的分类新技术―――boosting和bagging技术,为探究这些决策树分类算法与新技术在遥感影像分类方面的潜力,以中国华北地区MODIS250m分辨率影像进行了土地覆盖决策树分类试验与分析。研究结果表明决策树在满足充分训练样本的条件下,相对于传统方法如最大似然法(MLC)能明显提高分类精度,而在样本量不足下决策树分类表现差于MLC;并发现在单一决策树生成中,分类回归树CART算法表现较C4?5算法具有分类精度和树结构优势,分类精度的提高取决于树结构的合理构建与剪枝处理;另外在决策树CART中引入boosting技术,能明显提高那些较难识别类别的分类准确率18?5%到25?6%。

2003

0.0

2011

0.0

0.749

舒松, 余柏蒗, 吴健平, 等. 基于夜间灯光数据的城市建成区提取方法评价与应用[J]. 遥感技术与应用, 2011, 26(2): 169-175.

Shu

, Yu B

, Wu J

, et al. Methods for deriving urban built-up area using night-light data: Assessment and application[J]. Remote Sensing Technology and Application, 2011, 26(2): 169-175.

Night\|light data from the Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS) has been widely utilized to derive the urban built\|up areas.Four major methods,including Empirical Thresholding,Sudden Change Detection,Statistics\|assisted Thresholding,and TM\|assisted Thresholding,have been proposed to determine the threshold value for separating the illuminated urban areas from the dark background of rural areas in DMSP/OLS night\|light images.This paper makes a comprehensive assessment of those methods through a case study of Shanghai,China.The methods are implemented to extract the urban built\|up area for Shanghai using DMSP/OLS stable light data acquired in 2003.Then,the same threshold values obtained from 2003 DMSP/OLS night\|light data are applied respectively to the 2000 and 2006 DMSP/OLS data,resulting in a significant error in urban built\|up area detection.This analysis result suggests that the threshold value determined for a specific year cannot be extended and transferred to other years.The failure in temporal extensibility of threshold value means that an appropriate threshold value has to be determined for every year when a time series of DMSP/OLS nigh\|light data need to be processed.Therefore,the method that determines threshold value independent of reference data is more suitable for processing time series DMSP/OLS data.The Sudden Change Detection method does not require ancillary reference data and is the best choice of those methods considering the convenience,accuracy,and automated data processing,and it is then adopted to derive the urban built\|up areas of Shanghai from 2000 to 2006.

DMSP/OLS夜间灯光数据已被广泛应用于城市建成区的提取。目前主要存在4类提取方法:经验阈值法、突变检测法、统计数据法和较高分辨率影像数据空间比较法。以上海为例,在2000年、2003年、2006年夜间灯光数据的基础上,利用4种方法完成了城市建成区的提取。通过对不同年份数据提取结果的比较,证明了相同的灰度分割阈值对不同年份的夜间灯光数据中不存在通用性；在对2003年夜间灯光数据的提取中,4类方法所得结果精确度从高到低依次为统计数据法、突变检测法、经验阈值法和较高分辨率影像数据空间比较法,相对误差分别为1.3%、2.1%、5.1%和11.2%,在对4种方法的便捷性和可实现性进行分析与评价后,使用突变检测法完成了上海市2000~2006年城市建成区的提取。

... 综合TM图像、CART分类结果(像元值为DN_cart)和NTL(像元值为DN_ntl)3者叠加的统计结果,结合Milesi^[17]的经验阈值设定动态阈值,将CART 分类结果和NTL数据分别划分为3个层次(表2) ...

2009

1.613

0.0

... 扩张面积反映了某一时间段内城镇面积的增加总量,扩张速度为城镇面积年均增长量^[18,19],扩张强度是在某一段时间内年均城市面积增长比率^[20] ...

2007

0.0

林目轩, 师迎春, 陈秧分, 等. 长沙市区建设用地扩张的时空特征[J]. 地理研究, 2007, 26(2): 265-274.

Lin M

, Shi Y

, Chen Y

, et al. A study on spatial-temporal features of construction land expansion in Changsha urban area[J]. Geographical Research, 2007, 26(2): 265-274.

Construction land expansion is one of the notable features of urbanization.Analyzing the spatial-temporal features of construction land expansion can enhance the understanding of the nature of urbanization and provide with the grounds of decision-making for urban sustainable development. Based on the review of related research on construction land expansion at home and abroad, this paper takes Changsha urban area as an example, which is one of the most important cities in Central China.First, data on the variation of construction land in Changsha are obtained by employing the statistic and analytic function of GIS on the spatial data from remote sensing image data in different periods.Then the characteristics of construction land expansion are identified by using several methods in analyzing the spatial-temporal features of Changsha urban area from 1986 to 2004, which include expansion speed, expansion elasticity, fractal dimension, and rose map. The results obtained show that, during the last 20 years, the area of construction land has almost doubled, which is more obvious during the 2002-2004 period. Considering the relationship between the construction land area and population, the expansion speed of the construction land is far beyond that of the population, of which the erosion of arable land is most prominent. At present, the city sprawls by the way of low-density fill and the main direction of city expansion is NW270°-295° and SE165°-280°. Besides economic development and population growth, policy is also one of the important driving forces in construction land expansion. Especially in recent years, policy factor turns out to be the most determinant factor. In the future, the expansion mode of Changsha city will be that of mono-centric groups, and the direction of expansion will be more rational.

以不同时段的遥感影像资料为数据源,借助GIS技术对空间数据的统计及分析功能,提取出6个时期长沙市建成区变化的相关数据信息。在此基础上,利用扩张速度、扩张弹性、分维数、玫瑰图等方法分析长沙市区1986~2004年建设用地扩张的时空特征。结果表明,近20年来,长沙城市建设用地扩张了将近1倍,尤其是2002~2004年,建设用地扩张趋势最为明显。现阶段城市的扩张主要表现为低密度填充式的蔓延,城区扩张方向的主要方位是NW270°~295°和SE165°~180°。建设用地扩张的驱动因素除经济发展及人口增长因素之外,政策也是重要的影响因素,尤其是近年来政策因素的作用更为突出。未来长沙城市的空间发展将是多中心组团式的扩张,城市扩张方向日趋合理化。

... 扩张面积反映了某一时间段内城镇面积的增加总量,扩张速度为城镇面积年均增长量^[18,19],扩张强度是在某一段时间内年均城市面积增长比率^[20] ...

2012

0.0

0.4356

... 扩张面积反映了某一时间段内城镇面积的增加总量,扩张速度为城镇面积年均增长量^[18,19],扩张强度是在某一段时间内年均城市面积增长比率^[20] ...