EN中文

地球信息科学学报 >

2015 , Vol. 17 >Issue 12: 1497 - 1505

DOI: https://doi.org/10.3724/SP.J.1047.2015.01497

地球信息科学理论与方法

基于人类时空活动的城市土地利用分类研究

  • 鲁国珍 , 1 ,
  • 常晓猛 , 2*, * ,
  • 李清泉 1, 2 ,
  • 赵庆亮 3
展开
  • 1. 武汉大学 测绘遥感信息工程国家重点实验室,武汉 430079
  • 2. 深圳大学 空间信息智能感知与服务深圳市重点实验室,深圳 518060
  • 3. 北京市测绘设计研究院,北京 100038
*通讯作者：常晓猛（1986-）,男,江苏东台人,博士后,研究方向为地理社交网络、时空数据分析、智能交通与环境。E-mail:

作者简介：鲁国珍（1991-）,女,陕西西安人,硕士生,研究方向为时空数据分析与挖掘。E-mail:

收稿日期: 2015-10-10

  要求修回日期: 2015-10-28

  网络出版日期: 2015-12-20

基金资助

国家自然科学基金项目(41371377、41501486)

中国博士后基金项目(2015M572364)

CCF-腾讯犀牛鸟科研基金项目(CCF-TencentARG20150101)

城市空间信息工程北京市重点实验室经费重点项目(2014101)

深圳市科技研发基金项目(ZDSY20121019111146499、JCYJ20121019111128765)

收起

Land Use Classification Based on Massive Human-Activity Spatio-temporal Data

  • LU Guozhen , 1, 2 ,
  • CHANG Xiaomeng , 2, * ,
  • LI Qingquan 1, 2 ,
  • ZHAO Qingliang 3
Expand
  • 1. State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, China
  • 2. Shenzhen Key Laboratory of Spatial Smart Sensing and Services, Shenzhen University, Shenzhen 518060, China
  • 3. Beijing Institute of Surveying and Mapping, Beijing 100038, China
*Corresponding author: CHANG Xiaomeng, E-mail:

Received date: 2015-10-10

  Request revised date: 2015-10-28

  Online published: 2015-12-20

Copyright

《地球信息科学学报》编辑部 所有
Fold

摘要

准实时监测城市发展、掌握城市土地利用类型是日趋科学化、合理化进行城市规划的基本要求。随着信息通信技术、移动互联网技术、位置服务等的发展,海量的手机数据、浮动车数据、公交卡数据、社交网络数据等在内的人类时空活动信息为从“人”的角度动态实时感知城市土地利用、空间结构提供了可能。本文以深圳市为例,基于百万名QQ用户2013年的电子足迹数据,提出了不同类型的人类时空活动指数,以建立人类活动与城市地物间的对应关系;借鉴遥感影像不同波段记录各类地物在特定波谱区间辐射值的思想,生成各类人类时空活动指数波段图;并利用最大似然法对该“类高光谱影像”进行城市土地监督分类,获取城市的土地利用图。通过与深圳市规划图的对比验证,全体分类精度为72%。相较于传统基于“物”的遥感探测手段,基于“人”的城市感知更能反映城市内部相同地类的发展差异性。

关键词： 人类时空活动; 电子足迹; 社交网络数据; 土地利用; 土地覆盖; 分类

本文引用格式

鲁国珍 , 常晓猛 , 李清泉 , 赵庆亮 . 基于人类时空活动的城市土地利用分类研究[J]. 地球信息科学学报, 2015 , 17(12) : 1497 -1505 . DOI: 10.3724/SP.J.1047.2015.01497

Abstract

Macroscopically monitoring the status of urbanization and fast acquiring the land covers or land use in urban areas is essential for urban planning, management and scientific policy-making. The rapidly developing remote sensing technologies have been recognized as an essential approach to carry out this work because of their vital ability to capture the physical features of different land use, such as the spectral and textural properties. However, these technologies could not reveal the heterogeneity of urban development and differentiate the vitality in and among cities with the similar physical properties interpreted from remote sensing images. Human-activity based sensing technologies nowadays have been recognized as a promising alternative to resolve these problems. Spatio-temporal distribution of human activities could be derived from mobile phone records and smart card records stored in the public transportation systems, social media or social networking services (SNSs), and etc. They are good indicators for the social function of land use and urban vitality. We proposed types of indices to bridge the relationships between the intensities of human activities and land covers. Similar to the spectral bands of remote sensing images, more than thirty social bands were generated in this paper to describe the social characteristics of ground objects by aggregating and gridding human activities into pixels. According to the spectral profiles of eight land covers, a supervised classification approach was then applied to infer the land covers of the research area. Validation experiments were conducted in Shenzhen, China using a large-scale of people’s historical login information on Tencent QQ, which is the most popular SNS, during 2013. Results showed that the land cover of Shenzhen could be determined with a detection rate of 72% according to an urban planning map of Shenzhen. Compared to the classification results from remote sensing images, the human-activity based sensing technologies can obtain more detailed insight into the urban form, city skeleton, and the heterogeneity of development and vitality in different urban areas.

Key words： huamn activity; digital footprints; social network data; land use; land cover; classification

1 引言

随着中国社会生产力的发展与综合国力的增强,城市化进程正在迅猛地发展。城市数量的增多、建成区面积的扩大对准实时大面积监测城市发展提出了新的挑战;另一方面不断涌入的城市人口也对更加合理、和谐的城市居住环境提出了新的要求。因此,准实时掌握城市土地利用类型的变更成为科学、合理的城市规划及城市管理的基本要求。
遥感作为一种快速获取地表覆盖的重要手段,为城市的变化监测做出了巨大贡献。基于物理特性,尤其是光谱、纹理[1-4]的影像分类为大面积获取城市地类现状提供了便捷的手段;结合各类辅助信息,如上下文信息[1]、区域属性（大小、形状）[5-7]、专家知识[8]等的分类能大幅度提高分类精度,为城市变化监测的定量分析提供重要依据。同时,多空间、多时间、多光谱分辨率的卫星影像为城市从概貌到细节的获取提供了丰富的数据源。
然而,利用遥感技术进行变化监测的重要前提是遥感影像波谱反射辐射值的变化能很好地表现地表覆盖类型的变化,即不同地物类型在光谱上有不同的波谱反射特性[9]。这种基于物理特性的本质决定了遥感技术监测到的地表变化也仅仅是地表物理实体的变更,例如,建筑面积的不断扩大。然而,城市的发展程度并不仅仅体现于建成区的面积大小,人类活动的活跃程度是更具有经济属性的衡量指标;各个城市相同地类（尤其是建成区）内部在人类活动都表现为不同程度的差异性,尤其是对于开发时间不一致的各个城区,其发展差异性并不能通过遥感这种物理监测手段获取。新型城市化应该是人的城市化,而非人造的城市化[10]。因此,迫切需要基于人类活动的城市土地利用图来指导未来城市科学、合理的规划。
随着信息通信技术（Information Communication Technology,ICT）、基于位置服务（Location Based Services,LBS）等的发展,高精度、高频次的人类时空活动信息的获取成为可能[11-13],人人都是传感器的愿景正在实现[14-15]。海量、多源的人类时空活动数据为城市时空间行为研究提供了新的解决途径[15-17],基于“人”的城市感知成为城市研究的新角度[18]。Ratti最早论述了包含个体位置信息的人类活动数据对于城市分析的巨大潜力[19]。由于手机的广泛使用,人类活动的昼夜节律性通过手机通话记录（Call Detail Record,CDR）可清晰地反映,基于此构建的时间序列被广泛用于城市地类的推断[20-23]。不同于手机数据,GPS浮动车数据能够周期性的记录移动对象的位置信息,获得更加精细化的移动轨迹,不同时间、位置的上下车人数与城市地类间的关系也受到了学者的关注[24-25];公交IC卡所记录的居民出行行为则能在更大的空间尺度上分析城市的空间结构[26];另外,随着人们对于分享和表达诉求愿望的日益强烈,在社交网络上分享、上传包含时间戳与位置标签的文字、照片等也为人类活动研究提供了新的数据源[27-28]。相较于传统基于物理特征的简单地类划分,基于人类时空活动信息的城市地类提取更多的考虑了社会经济等属性。
然而,不同的人类时空活动数据也在不同程度上存在着缺点。首先,部分数据的定位精度不够高。例如,手机通话所记录的时空活动信息由基站记录[14, 29-30],因此,定位精度只能达到基站级别,且取决于基站的密集程度。其次,时空活动信息仅在事件被用户触发时才被动记录,例如,Flickr上收集的带有位置标签的照片需要用户主动上传位置信息[15];再次,时空数据的分布具有明显的稀疏性,例如,签到数据更多的分布于商业区而非住宅区。

2 研究区域概况及数据源

本文选择深圳市为研究区,其地处广东省南部,珠江三角洲东岸,与香港一水之隔,东临大亚湾和大鹏湾,西濒珠江口和伶仃洋,南隔深圳河与香港相连,北部与东莞、惠州接壤。深圳市总面积2020.5 km2,自1980年成为特区以来,由一个人口33万的沿海渔村小镇发展成为中国最大的移民城市,截止2014年常住人口约1077万人[32]。伴随着特区的城市化进程,曾经的山地、海洋等大面积转化为建成区,这种土地类型的变更随着城市人口的扩张仍在进行。
本文采用的人类时空活动数据来自腾讯QQ-中国用户量最大的即时通讯服务提供商。截止到2014年11月,QQ的月活跃用户已经达到8.29亿,同时在线人数超过2亿[33]。实验从8.29亿活跃用户中随机抽取了34 047 101匿名用户（其中深圳市用户4 053 216）,并收集了这些用户在2013年共 548 433 983条轨迹点信息。为了保护用户的隐私,对所有用户的ID进行了多次加密,移动轨迹被进行了一定程度的空间扰动。

3 基于人类电子足迹数据的土地利用类型获取方法

为了提取城市的土地利用类型,本文首先定义一系列指数来建立不同地类与人类活动之间的对应关系,通过统计格网内的指数值来生成相应的指数波段,并构建基于“人”感知的“类高光谱影像”,借助于不同地类在“类高光谱影像”上的不同响应值,实现基于人类活动的土地利用类型提取。
实验处理流程（图1）主要包括以下步骤：（1）选择恰当的分辨率将时空活动数据格网化;（2）定义一系列指数定量描述人类活动,生成基于人类时空活动的“类高光谱影像”;（3）选取不同地物的样本并统计地物的波谱响应曲线,定量分析地物的可分离性;（4）利用监督分类获得基于人类时空活动的土地利用图;（5）结果分析与精度评价。
Fig. 1 Framework of urban land user classification using human-activity tracking data

图1 基于人类时空活动提取土地利用类型的流程图

3.1 时空活动数据的格网化

实验的原始数据为每个匿名用户在一年内登录的位置信息（经度、纬度值）,为了便于从不同角度提取特征,首先将研究区域划分为一定大小的格网。考虑GPS的定位精度,实验将深圳市划分为0.6 s的格网,约为18 m×18 m,每个格网可看做遥感影像中的一个像元,整张影像覆盖深圳地区。每个像元的DN值反映了相应指数描述下人类活动的强度信息。

3.2 “类高光谱影像”合成

3.2.1 人类活动指数的定义
鉴于人们日常活动最频繁的地点一般为住宅或工作地,因此,本文假设用户访问频次最高的位置（Top 2）（一般为住宅或办公用地）对应的格网为该用户的归属地。首先统计研究区域内每个格网中落入的最高频次（Top 2）的用户数量,即为该格网内的本地用户量,并用本地用户量指数NLR（Number of Local Residents）来定量描述本地用户的分布如式(1)。
NL R i = ∑ p = 1 P b p   ,     b p = 1 ,           i ∈ { g p }   0 ,         i ∉ { g p }         (1)
式中, { g p }   是用户 p 历史登录频次最高的2个位置; P 是研究区域内的本地用户总量;同时,由于NLR与当地的人口密度直接相关,因此定义指数DLR（Density of Local Residents）来反映本地用户的密度(式2)。
式中, A m 是区域 m 的所有格网点; Π m 是区域 m 内本地用户总量。
除了以上2个指数,本文还定义了3类吸引力指数AI（Attractiveness Indicators）,分别为访问人数指数NV（Number of Visitors）、访问强度指数TIV（Total Intensity of Visits）和平均访问强度指数AIV（Average Intensity of Visits）;并根据地域的不同,对每一类指数分别从全体用户、本地用户、外地用户3个方面进行统计分析。具体的指数描述如表1。
Tab. 1 The description of human activity bands

表1 人类活动指数描述

指数 波段 波段全称 公式 统计方式 指数含义
NV NWV Number of Whole Visitors 反映格网被访问的人数,用以提取诸如博物馆、公园、旅游景点等访问人数很高的区域
NLV Number of Local Visitors
NOV Number of Outside Visitors
TIV TDW Total Visit Days of Whole Visitors 用户访问总天数 反映格网被访问的人次,用于提取诸如工作地、住宅地等被频繁访问的区域
TDL Total Visit Days of Local Visitors
TDO Total Visit Days of Outside Visitors
TTW Total Visit Times of Whole Visitors 用户访问总次数
TTL Total Visit Times of Local Visitors
TTO Total Visit Times of Outside Visitors
AIV ADW Average Visit Days of Whole Visitors ADW= TDWNWV 用户访问总天数 访问人次与访问人数的比值
ADL Average Visit Days of Local Visitors ADL= TDLNLV
ADO Average Visit Days of Outside Visitors ADO= TDONOV
ATW Average Visit Times of Whole Visitors ATW= TTWNWV 用户访问总次数
ATL Average Visit Times of Local Visitors ATL= TTLNLV
ATO Average Visit Times of Outside Visitors ATO= TTONOV
ADWPLR Averge Visit Days of Whole Visitors by Per-capita Local Resident ADWPLR= TDWNLR 用户访问总天数 访问人次与本地用户密度的比值
ADLPLR Averge Visit Days of Local Visitors by Per-capita Local Resident ADLPLR= TDLNLR
ADOPLR Averge Visit Days of Outside Visitors by Per-capita Local Resident ADOPLR= TDONLR
ATWPLR Average Visit Times of Whole Visitors by Per-capita Local Resident ATWPLR= TTWNLR 用户访问总次数
ATLPLR Average Visit Times of Local Visitors by Per-capita Local Resident ATLPLR= TTLNLR
ATOPLR Average Visit Times of Outside Visitors by Per-capita Local Resident ATOPLR= TTONLR
3.2.2 合成“类高光谱影像”波段上述人类活动指数定义生成相应的波段,获得基于人类时空活动数据的类高光谱影像,以反映人类时空活动相应于不同指数的特征。类似地,通过任意3个波段的假彩色合成影像也有助于目视解译。如图2所示,城市的街区和主干路已能清晰地分辨出来。
Fig. 2 Example of three bands and their false color composite image derived from human-activity sensing images

图2 “类高光谱影像”的示例波段及合成的假彩色影像

本文通过对波段进行相关性分析,共选取了以上22个波段中的9个波段进行分析,分别是NLR（Band I）、NLV（Band II）、NOV（Band III）、TDL（Band IV）、TDO（Band V）、ADL（Band VI）、ADO（Band VII）、ADLPLR（Band VIII）和ADOPLR（Band IX）。

3.3 “类高光谱影像”地物波谱响应分析

3.3.1 基于人类时空活动数据的土地分类系统
利用影像进行地类推断最早被广泛应用于航空影像[34],然而,由于分类系统不统一,增加了大量的重复性劳动。Anderson于1976年最早对基于遥感影像的分类系统进行了标准化工作,将其划分为9个一级类,37个二级类[35]。这9个一级类分别为：建成区（Urban or Built-up Land）、农业用地（Agricultural Land）、牧草地（Rangeland）、林地（Forest Land）、水体（Water）、湿地（Wetland）、裸地（Barren Land）、冻土地（Tundra）、积雪地（Perennial Snow or Ice）。目前,遥感影像的分类系统大多采用其中的部分或根据实际需求调整而得[36]。
由于该分类系统已被广泛认可和应用,本文的分类工作将参考该分类系统进行;加上人类活动在山地、水库、湖泊等人迹罕至区域几乎无数据,因此本文的分类工作重点围绕人类活动相对频繁的建成区展开。首先,将建成区划分为日常工作涉及较多的建筑群区域（Buildings）、交通用地（Transportation Areas）,以及更趋于休闲、娱乐的市内绿地（Urban Green Land）、市内水体（Urban Water Surface）;其次,由于人类活动在相同地类的频繁程度存在差异,进一步将建筑群（Buildings）划分为活动密集型（Buildings with dense-activity）和活动稀疏型（Buildings with sparse-activity）2类,对于交通用地则根据人流量的大小划分为一级道路（Main Road）和二级道路（Secondary Road）。至于建成区外少量极稀疏的数据点区域,简单划分为市外绿地（Field Vegetation）、市外水体（Field Water Surface）。
3.3.2 波谱响应曲线分析
为了定量分析“类高光谱影像”对于地物的可分离性,实验选取了研究区域内包括活动密集型建筑群（Buildings with dense-activity）、活动稀疏型建筑群（Buildings with sparse-activity）、一级道路（Main Road Surface）、二级道路（Secondary Road Surface）、市内绿地（Urban Green Land）、市外绿地（Field Vegetation）、市内水体（Urban Water Surface）、市外水体（Field Water Surface）8类地物的样本,统计了各类地物在不同波段的均值,获得了地物关于“类高光谱影像”的响应曲线,如图3所示。
Fig. 3 Spectral profiles derived from human sensing images of nine major land cover types in Shenzhen area

图3 深圳地区的“类高光谱影像”的波谱响应曲线

由图3可看出,市外水体、市外绿地在各个波段的响应值都极小,而活动密集型建筑群和一级道路在各个波段的响应值都很大,尤其是Band I（NLR）、Band VI（ADL）。对于大部分波段,活动密集型建筑群的响应值都大于一级道路,差异最大的波段为Band IV（TDL）和Band V（TDO）;然而,Band VII（ADO）一级道路的响应值比活动密集型建筑群大。活动稀疏型建筑群、二级道路、市内绿地、市内水体的响应值介于一级道路与市外绿地之间;其中,二级道路与市内绿地响应曲线的趋势类似,二级道路的响应值均大于市内绿地。相较于二级道路和市内绿地,活动稀疏型建筑群在Band I（NLR）、Band III（NOV）、 Band VII （ADO）值最大,在Band IV（TDL）、Band VIII（ADLPLR）值则最小。
上述分析可看出,活动密集型建筑群、一级道路具有较高DN值,而市外绿地、市外水体具有较低DN值,二者区分明显;根据Band IV（TDL）（或 Band V）、Band VII（ADO）的特点,活动密集型建筑群与一级道路有可能区分开;尽管市外绿地、市外水体在Band VII（ADO）到Band IX（ADOPLR）的各个波段均有少许差别,但二者依然很难分开。根据Band I（NLR）、Band III（NOV）和Band VII（ADO）,首先可区分活动稀疏型建筑群与其他地物,而二级道路在大部分波段都具有较高的响应值,因此可与市内绿地区分开。

4 “类高光谱影像”土地利用及精度 评价

4.1 基于“类高光谱影像”的土地利用图

本文利用传统的监督分类方法（最大似然法）对基于人类时空活动的“类高光谱影像”进行分类。为了降低人工选样对分类结果的干扰,实验从深圳市土地规划图中随机选取1000个建成区、道路、绿地3类地物的多边形（约占总体的1%）,其中,一半作为监督分类的样本,一半作为分类精度评价的真实ROI。
为了提高分类结果的可靠性,实验共进行了10次,分类结果如图4所示。活动密集型建筑群与活动稀疏型建筑群的主要差异为人类活动的频繁程度不同。图4(c)为深圳的罗湖区、福田区交界区域,属于深圳开发较早的城区,是深圳主要的金融区与商业中心区,因此在活动频次、活动人次上都表现为活动密集型的区域;图4(b)为深圳的宝安区,近几年才开始开发,尽管高楼林立,但活动强度仍明显弱于老城区,属于活动稀疏型建筑群。对于道路的分类,在一些高热点区域,道路和活动密集型建筑群会出现混淆,如图4(c)的最右边部分建筑群被分为一级道路,主要原因在于该地区包含有大剧院、国贸等人流量大、访问频次低的POI（Point of Interest）,因此,从人类时空活动的角度出发,这些区域与一级道路的特征更为接近;对于一般的城区,街区内的道路可较为完整、准确地提取（图4(b)）。同时,公园、高尔夫球场等市内绿地也能获得,但是与市外道路被混为一类,原因仍在二者均表现为相对较高的人流量和相对较低的访问频次等类似特征,区分这2类区域是下一步的工作之一。
Fig. 4 Result of land use classification in Shenzhen

图4 深圳市的土地利用分类结果图

4.2 精度评价

本文将深圳市的规划图视为真实ROI,经验证实验的全体分类精度为72.44%,kappa系数0.621。
通过分析混淆矩阵（表2）、生产者精度和用户精度（图5）可看出,绿地具有最高的生产者精度（除去未分类）,主要是因为绿地在“类高光谱影像”上的响应值要比建成区、道路低很多,而又明显强于水体等几乎无响应值的未分类区域,因此,更易于与其他地物区分,分类精度较高。另外,道路的用户精度最高（图5）,说明基于人类时空活动的“类高光谱影像”对于探测道路具有很大优势;但是道路的生产者精度不到60%,建成区的生产者精度稍高（68%）,这主要是建成区与道路在热点区域具有相似特征,即二者在人类时空活动强度方面表现一致,因此无法区分;图4 (c)中的混淆区域即是这一现象的说明。
Fig. 5 Producer’s and user’s accuracies of the classification

图5 分类结果的用户精度、生产者精度

Tab. 2 Confusion matrix of the classification

表2 分类结果的混淆矩阵

建筑群 道路 市内绿地 其他
建筑群 0.6807 0.2037 0.0903 0
道路 0.1171 0.5955 0.0602 0
市内绿地 0.1431 0.1913 0.7398 0.0024
其他 0.0591 0.0095 0.1097 0.9976

5 结论与展望

近年来,不断加剧的城市扩张推进着中国各城市的城市化进程,土地利用类型的频繁变更对实时监测城市发展提出了新的挑战,迅速掌握城市土地利用类型成为城市规划、城市管理的基本要求。ICT、LBS技术等的发展,使得高精度、高频次的人类时空活动信息的获取成为可能,也为基于人类活动的土地利用类型提取带来新的解决方法。
本文提出一系列指数用以定量描述人类时空活动,并基于相应的指数波段生成“类高光谱影像”,通过定量分析地物的可分离性,利用一种监督分类方法获得了深圳市的土地利用图,并以深圳市规划图为真实ROI求得全体分类精度为72.44%。本研究首次将基于“物”的遥感探测思想应用于基于“人”的城市感知领域,通过不同的指数来量化人类时空活动特征,并转化为“类高光谱影像”来综合利用各个指数特征实现地物区分。
当然,关于此方面的研究还在探索中,本文的工作难免存在一些不足,尤其是还未能将分类结果细化到更具有应用价值的城市功能区中,未来将主要围绕以下2个方面加以完善：（1）加入包含时间特征的指数以优化人类时空活动的描述。包含日内时间变化的指数波段将是区分工作区、住宅区的有力依据;而对于混为一类的市内绿地与市外道路,包含日间时间变化的指数,尤其是工作日与非工作日的不同时间描述,能为二者的区分提供重要参考。（2）融合遥感影像与“类高光谱影像”,利用遥感影像获取山地、水库、湖泊等地类,解决人类时空活动数据在这些人烟罕至地缺乏数据的弊端,综合解决基于物理特征和基于功能区的城市土地利用提取问题。

The authors have declared that no competing interests exist.

参考文献

原文顺序 | 文献年度倒序 | 文中引用次数倒序
[1]
Gong P, Howarth P.The use of structural information for improving land-cover classification accuracies at the rural-urban fringe[J]. Photogrammetric engineering and remote sensing, 1990,56(1):67-73.A methodology for incorporating structural information into conventional classification procedures is described. The technique is based on the use of an edge-density image which is generated using a Laplacian operator. This image is included in a Mahalanobis classifier as an additional band of data. The method is particularly designed for higher spatial resolution data. It has been tested using...

DOI

[2]
Shaban M, Dikshit O.Improvement of classification in urban areas by the use of textural features: The case study of Lucknow city, Uttar Pradesh[J]. International Journal of Remote Sensing, 2001,22(4):565-593.Investigations have been carried out for digital spectral and textural classification of an Indian urban environment using SPOT images with grey level co-occurrence matrix (GLCM), grey level difference histogram (GLDH), and sum and difference histogram (SADH) approaches. The results indicate that a combination of texture and spectral features significantly improves the classification accuracy compared with classification with pure spectral features only. This improvement is about 9% and 17% for an addition of one and two texture features, respectively. GLDH and SADH give statistically similar results to GLCM, and take less computing time than GLCM. Conventional separability measures like transformed divergence, Bhattacharya distance, etc. are not effective in feature selection when classification is carried out with spectral and texture features. An alternative approach using simple statistics such as average coefficient of variation, skewness, and kurtosis and correlation amongst feature sets has shown greater feature selection potential when a combination of spectral and texture features is used.

DOI

[3]
Fisher P.The pixel: A snare and a delusion[J]. International Journal of Remote Sensing, 1997,18(3):679-685.The pixel is an explicit feature of remotely-sensed imagery, and a primary concept of the raster GIS (Geographical Information System) which is the usual vehicle for integration. This Letter addresses the underlying spatial conceptualization of the pixel, which is the parallel of the grid cell in spatial analysis, and the regular grid in sampling. It is argued that integration of remote sensing and GIS can only possibly advance if we develop methods to address the conceptual short-comings of the pixel as a spatial entity, and stop pretending that it is a true geographical object. Three major strands of research which address this issue are highlighted, including mixture modelling, geostatistics and fuzzy classification.

DOI

[4]
Lu D, Weng Q.Use of impervious surface in urban land-use classification[J]. Remote Sensing of Environment, 2006,102(1):146-160.

[5]
Wu S S, Qiu X, Usery E L, et al.Using geometrical, textural, and contextual information of land parcels for classification of detailed urban land use[J]. Annals of the Association of American Geographers, 2009,99(1):76-98.Detailed urban land use data are important to government officials, researchers, and businesspeople for a variety of purposes. This article presents an approach to classifying detailed urban land use based on geometrical, textural, and contextual information of land parcels. An area of 6 by 14 km in Austin, Texas, with land parcel boundaries delineated by the Travis Central Appraisal District of Travis County, Texas, is tested for the approach. We derive fifty parcel attributes from relevant geographic information system (GIS) and remote sensing data and use them to discriminate among nine urban land uses: single family, multifamily, commercial, office, industrial, civic, open space, transportation, and undeveloped. Half of the 33,025 parcels in the study area are used as training data for land use classification and the other half are used as testing data for accuracy assessment. The best result with a decision tree classification algorithm has an overall accuracy of 96 percent and a kappa coefficient of 0.78, and two naive, baseline models based on the majority rule and the spatial autocorrelation rule have overall accuracy of 89 percent and 79 percent, respectively. The algorithm is relatively good at classifying single-family, multifamily, commercial, open space, and undeveloped land uses and relatively poor at classifying office, industrial, civic, and transportation land uses. The most important attributes for land use classification are the geometrical attributes, particularly those related to building areas. Next are the contextual attributes, particularly those relevant to the spatial relationship between buildings, then the textural attributes, particularly the semivariance texture statistic from 0.61-m resolution images.

DOI

[6]
Hu S, Wang L.Automated urban land-use classification with remote sensing[J]. International Journal of Remote Sensing, 2013,34(3):790-803.Urban land-use information plays a key role in a wide variety of planning and environmental management processes. The purpose of this study was to develop an automatic method for classifying detailed urban land-use classes with remote-sensing data. Seven land-use parcel attributes, derived from relevant remote-sensing data, were incorporated for classifying four land-use classes, namely office, industrial, civic, and transportation, which were reported as the most difficult ones to classify from previous studies. An experiment was carried out in a study site in Austin, Texas. An overall accuracy of 61.68% and a kappa coefficient of 0.54 were achieved with a decision tree method. Building area and building height turned out to be the most influential factors among all the adopted variables. In addition, the variable of floor area ratio played the second dominant role among the seven variables, demonstrating that synthesized horizontal and vertical properties of buildings and their relevant spatial characteristics are important in differentiating the four classes.

DOI

[7]
De Wit A, Clevers J.Efficiency and accuracy of per-field classification for operational crop mapping[J]. International journal of remote sensing, 2004,25(20):4091-4112.A crop map of The Netherlands was created using a methodology that integrates multi-temporal and multi-sensor satellite imagery, statistical data on crop area and parcel boundaries from a 1 : 10 000 digital topographic map. In the first phase a crop field database was created by extracting static parcel boundaries from the digital topographic map and by adding dynamic crop boundaries using on-screen digitizing. In the next phase the crop type was determined from the spectral and phenological properties of each field. The resulting crop map has an accuracy larger than 80% for most individual crops and an overall accuracy of 90%. By comparing cost and man-hours it was demonstrated that per-field classification is more efficient than per-pixel classification and decreased the effort for classification from 1500 to 500 man-hours, but the effort for creating the crop field database was estimated at 2300 man-hours. The use of image segmentation techniques for deriving the crop field database was discussed. It was concluded that image segmentation cannot replace the use of a large-scale topographic map but, in the future, image segmentation may be used to map the dynamic crop boundaries within the topographic parcels.

DOI

[8]
Platt R V, Rapoza L.An evaluation of an object-oriented paradigm for land use/land cover classification[J]. The Professional Geographer, 2008,60(1):87-100.Object-oriented image classification has tremendous potential to improve classification accuracies of land use and land cover (LULC), yet its benefits have only been minimally tested in peer-reviewed studies. We aim to quantify the benefits of an object-oriented method over a traditional pixel-based method for the mixed urban-uburban-gricultural landscape surrounding Gettysburg, Pennsylvania. To do so, we compared a traditional pixel-based classification using maximum likelihood to the object-oriented image classification paradigm embedded in eCognition Professional 4.0 software. This object-oriented paradigm has at least four components not typically used in pixel-based classification: (1) the segmentation procedure, (2) nearest neighbor classifier, (3) the integration of expert knowledge, and (4) feature space optimization. We evaluated each of these components individually to determine the source of any improvement in classification accuracy. We found that the combination of segmentation into image objects, the nearest neighbor classifier, and integration of expert knowledge yields substantially improved classification accuracy for the scene compared to a traditional pixel-based method. However, with the exception of feature space optimization, little or no improvement in classification accuracy is achieved by each of these strategies individually.

DOI

[9]
祝锦霞. 高分辨率遥感影像变化检测的关键技术研究[D]. 杭州:浙江大学,2011.

[10]
从鬼城看中国城市化前景.

[11]
陆锋,刘康,陈洁.大数据时代的人类移动性研究[J].地球信息科学学报,2014,16(5):665-672.<p>人类个体/群体移动特征是多学科共同关注的研究主题。移动定位、无线通讯和移动互联网技术的快速发展使得获取大规模、长时间序列、精细时空粒度的个体移动轨迹和相互作用定量化成为可能。同时，地理信息科学、统计物理学、复杂网络科学和计算机科学等多学科交叉也为人类移动性研究的定量化提供了有力支撑。本文首先系统总结了大数据时代开展人类移动性研究的多源异构数据基础和多学科研究方法，然后将人类移动性研究归纳为面向人和面向地理空间两大方向。面向人的研究侧重探索人类移动特性的统计规律，并建立模型解释相应的动力学机制，或分析人类活动模式，并预测出行或活动；面向地理空间的研究侧重从地理视角分析人类群体在地理空间中的移动，探索宏观活动和地理空间的交互特征。围绕这两大方向，本文评述了人类移动性的研究进展和存在问题，认为人类移动性研究在数据稀疏性、数据偏斜影响与处理、多源异构数据挖掘、机器学习方法等方面依然面临挑战，对多学科研究方法的交叉与融合提出了更高要求。</p>

DOI

[12]
陆锋,张恒才.大数据与广义GIS[J].武汉大学学报(信息科学版),2014,39(6):645-654.<b>目的</b>普适计算基础设施和数据处理技术的发展催生了大数据概念,而大数据时空粒度的不断细化加速了地理空间信息的泛化过程。阐述了大数据时代地理空间信息泛化的显著特征,进而提出GIS概念广义化的迫切需求,从数据采集与整理、数据管理与集成、数据分析与计算三个方面分析了广义GIS所面临的技术挑战,重点探讨了互联网蕴含地理空间数据采集、移动对象数据库和异构动态数据管理、移动对象轨迹数据挖掘、复杂网络分析等方面的研究进展与存在的问题,并展望了广义GIS时代地理计算与城市计算、社会计算的融合趋势。

DOI

[13]
李清泉,李德仁.大数据GIS[J].武汉大学学报(信息科学版), 2014,39(6):641-644,666.<b>目的</b>大数据不仅使世界认识到数据的重要性,更引发了许多行业从根本上的变革。大数据时代也对GIS提出了诸多挑战,如海量、多源、异构数据的存储与管理以及面对大量噪音的价值挖掘等。作为空间数据管理、分析以及可视化的重要工具,为适应大数据的需求,GIS必须在大数据时代做出改变和调整。针对大数据的几个“V”特性分析了传统GIS所受到的挑战,在前期相关研究的基础上,从GIS空间数据管理、空间数据分析以及可视化三方面进一步阐述了大数据GIS应具有的特征。

DOI

[14]
Goodchild M F.Citizens as sensors: The world of volunteered geography[J]. GeoJournal, 2007,69(4):211-21.In recent months there has been an explosion of interest in using the Web to create, assemble, and disseminate geographic information provided voluntarily by individuals. Sites such as Wikimapia and OpenStreetMap are empowering citizens to create a global patchwork of geographic information, while Google Earth and other virtual globes are encouraging volunteers to develop interesting applications using their own data. I review this phenomenon, and examine associated issues: what drives people to do this, how accurate are the results, will they threaten individual privacy, and how can they augment more conventional sources? I compare this new phenomenon to more traditional citizen science and the role of the amateur in geographic observation.

DOI

[15]
Song C, Qu Z, Blumm N, Barabási A-L.Limits of predictability in human mobility[J]. Science, 2010,327(5968):1018-1021.Abstract A range of applications, from predicting the spread of human and electronic viruses to city planning and resource management in mobile communications, depend on our ability to foresee the whereabouts and mobility of individuals, raising a fundamental question: To what degree is human behavior predictable? Here we explore the limits of predictability in human dynamics by studying the mobility patterns of anonymized mobile phone users. By measuring the entropy of each individual's trajectory, we find a 93% potential predictability in user mobility across the whole user base. Despite the significant differences in the travel patterns, we find a remarkable lack of variability in predictability, which is largely independent of the distance users cover on a regular basis.

DOI PMID

[16]
柴彦威,塔娜.中国时空间行为研究进展[J].地理科学进展,2013,32(9):1362-1373.自时间地理学和活动分析法引入中国以来的近20年间,时空间行为研究已经成为中国城市地理学的重要领域。中国时空间行为研究关注城市空间重构的描述与解释,试图从行为角度解释中国城市社会转型,强调转型期中国城市空间与居民个体行为之间的互动关系,重视日常生活、生活质量、社会公正、低碳社会、智慧城市等热点问题,探索在城市交通、旅游和城市规划等领域中的实践应用。中国时空间行为研究已经形成了以解读城市转型为目标、以规划应用为导向的鲜明特点,为理解中国城市制度与空间转型背景下人类行为模式的复杂性和多样性提供了一个全新的视角。但是,中国时空间行为研究依然面临着理论发展滞后、实践应用需要突破等挑战。本文是对时空行为研究近年来发展的综述性文章,从数据采集与分析方法演进、实证研究与规划应用进展等方面回顾了近20年来中国城市时空间行为研究的最新进展,致力于推动不同学科领域之间的交流和时空间行为研究自身的发展。

DOI

[17]
柴彦威,沈洁.基于活动分析法的人类空间行为研究[J].地理科学,2008,28(5):594-600.居民日常活动空间和城市日常活动系统是人类空间行为研究中的重要内容,直接反映行为空间形成机制、分布特征及其与实体空间的相互关系,并为城市社会生活及其空间体系研究提供了重要的微观视角。通过行为空间研究中若干关键问题进行探讨,认为基于日常活动空间的活动分析法能够很好的整合和处理这些问题,在对活动分析法相关概念和几个重要方法介绍的基础上,尝试提出基于活动分析法的人类空间行为研究的思路与框架。

[18]
刘瑜,肖昱,高松,等. 基于位置感知设备的人类移动研究综述[J].地理与地理信息科学,2011,27(4):8-13,31.每个人在地理空间内的移动看似 随机而没有规律,然而一个较大规模人群的移动却隐藏着特定的模式。为了研究某些地理问题,如交通、疾病传播等,可以从个体行为出发,在地理信息系统的支持 下,发现人类移动模式,并构筑基于个体的模拟模型,从而建立微观和宏观之间的桥梁,并支持相应的决策过程。信息通讯技术的发展,一方面改变了人们的空间行 为模式,另一方面使得基于位置感知设备获取海量人类移动数据成为可能。近年来,上述研究一直是地理信息科学及相关领域的热点,该文对此进行了总结和评述。

[19]
Ratti C, Williams S, Frenchman D, et al.Mobile landscapes: Using location data from cell phones for urban analysis[J]. Environment and Planning B Planning and Design, 2006,33(5):727.The technology for determining the geographic location of cell phones and other handheld devices is becoming increasingly available. It is opening the way to a wide range of applications, collectively referred to as location-based services (LBS), that are primarily aimed at individual users. However, if deployed to retrieve aggregated data in cities, LBS could become a powerful tool for urban analysis. In this paper we aim to review and introduce the potential of this technology to the urban planning community. In addition, we present the ‘Mobile Landscapes’ project: an application in the metropolitan area of Milan, Italy, based on the geographical mapping of cell phone usage at different times of the day. The results enable a graphic representation of the intensity of urban activities and their evolution through space and time. Finally, a number of future applications are discussed and their potential for urban studies and planning is assessed.

DOI

[20]
Soto V, Frías-Martínez E.Automated land use identification using cell-phone records[C]. Proceedings of the 3rd ACM international workshop on MobiArch, 2011.

[21]
Soto V, Frias-Martinez E.Robust land use characterization of urban landscapes using cell phone data[C]. Proceedings of the 1st Workshop on Pervasive Urban Applications, in conjunction with 9th International Conference on Pervasive Computing, 2011.

[22]
Toole J L, Ulm M, González M C, et al.Inferring land use from mobile phone activity[C]. Proceedings of the ACM SIGKDD international workshop on urban computing, 2012.

[23]
Pei T, Sobolevsky S, Ratti C, et al.A new insight into land use classification based on aggregated mobile phone data[J]. International Journal of Geographical Information Science, 2014,28(9):1988-2007.Land-use classification is essential for urban planning. Urban land-use types can be differentiated either by their physical characteristics (such as reflectivity and texture) or social functions. Remote sensing techniques have been recognized as a vital method for urban land-use classification because of their ability to capture the physical characteristics of land use. Although significant progress has been achieved in remote sensing methods designed for urban land-use classification, most techniques focus on physical characteristics, whereas knowledge of social functions is not adequately used. Owing to the wide usage of mobile phones, the activities of residents, which can be retrieved from the mobile phone data, can be determined in order to indicate the social function of land use. This could bring about the opportunity to derive land-use information from mobile phone data. To verify the application of this new data source to urban land-use classification, we first construct a vector of aggregated mobile phone data to characterize land-use types. This vector is composed of two aspects: the normalized hourly call volume and the total call volume. A semi-supervised fuzzy c-means clustering approach is then applied to infer the land-use types. The method is validated using mobile phone data collected in Singapore. Land use is determined with a detection rate of 58.03%. An analysis of the land-use classification results shows that the detection rate decreases as the heterogeneity of land use increases, and increases as the density of cell phone towers increases.

DOI

[24]
Qi G, Li X, Li S, et al. Measuring social functions of city regions from large-scale taxi behaviors[C].2011 IEEE International Conference on Pervasive Computing and Communications Workshops (PERCOM Workshops), 2011.

[25]
Liu Y, Wang F, Xiao Y, et al.Urban land uses and traffic source-sink areas: Evidence from GPS-enabled taxi data in Shanghai[J]. Landscape and Urban Planning, 2012,106(1):73-87.Most of the existing literature focuses on estimating traffic or explaining trip lengths from land use. This research attempts to reveal intraurban land use variations from traffic patterns. Using a seven-day taxi trajectory data set collected in Shanghai, we investigate the temporal variations of both pick-ups and drop-offs, and their association with different land use features. Based on the balance between the numbers of drop-offs and pick-ups and its distinctive temporal patterns, the study area is classified into six traffic ‘source-sink’ areas. These areas are closely associated with various land use types (commercial, industrial, residential, institutional and recreational) as well as land use intensity. The study shows that human mobility data from location aware devices provide us an opportunity to derive urban land use information in a timely fashion, and help urban planners and policy makers in mitigating traffic, planning for public services and resources, and other purposes.

DOI

[26]
Han H, Yu X, Long Y. Discovering functional zones using bus smart card data and points of interest in Beijing[EB/OL]. .

[27]
Frias-Martinez V, Soto V, Hohwald H, et al.Characterizing urban landscapes using geolocated tweets[C]. Proceedings of the 2012 ASE/IEEE International Conference on Social Computing and 2012 ASE/IEEE International Conference on Privacy, Security, Risk and Trust, 2012.

[28]
Hollenstein L, Purves R.Exploring place through user-generated content: Using Flickr tags to describe city cores[J]. Journal of Spatial Information Science, 2015(1):21-48.Terms used to describe city centers, such asDowntown, are key concepts in every-day or vernacular language. Here, we explore such language by harvesting georeferenced and tagged metadata associated with 8 million Flickr images and thus consider how large numbers of people name city core areas. The nature of errors and imprecision in tagging and georeferencing are quantified, and automatically generated precision measures appear to mirror errors in the positioning of images. Users seek to ascribe appropriate semantics to images, though bulk-uploading and bulk-tagging may introduce bias. Between 0.5--2% of tags associated with georeferenced images analyzed describe city core areas generically, while 70 % of all georeferenced images analyzed include specific place name tags, with place names at the granularity of city names being by far the most common. Using Flickr metadata, it is possible not only to describe the use of the term Downtown across the USA, but also to explore the borders of city center neighborhoods at the level of individual cities, whilst accounting for bias by the use of tag profiles.

DOI

[29]
Gonzalez M C, Hidalgo C A, Barabasi A-L.Understanding individual human mobility patterns[J]. Nature, 2008,453(7196):779-82.Despite their importance for urban planning, traffic forecasting, and the spread of biological and mobile viruses, our understanding of the basic laws governing human motion remains limited thanks to the lack of tools to monitor the time resolved location of individuals. Here we study the trajectory of 100,000 anonymized mobile phone users whose position is tracked for a six month period. We find that in contrast with the random trajectories predicted by the prevailing Levy flight and random walk models, human trajectories show a high degree of temporal and spatial regularity, each individual being characterized by a time independent characteristic length scale and a significant probability to return to a few highly frequented locations. After correcting for differences in travel distances and the inherent anisotropy of each trajectory, the individual travel patterns collapse into a single spatial probability distribution, indicating that despite the diversity of their travel history, humans follow simple reproducible patterns. This inherent similarity in travel patterns could impact all phenomena driven by human mobility, from epidemic prevention to emergency response, urban planning and agent based modeling.

DOI PMID

[30]
Calabrese F, Di Lorenzo G, Liu L, et al.Estimating origin-destination flows using mobile phone location data[J]. IEEE Pervasive Computing, 2011:36-44.Using an algorithm to analyze opportunistically collected mobile phone location data, the authors estimate weekday and weekend travel patterns of a large metropolitan area with high accuracy.

DOI

[31]
微信光芒万丈,拥有8亿月活跃用户的QQ到底在想什么?.

[32]
2014年深圳市常住人口中户籍与非户籍人口数.

[33]
腾讯QQ同时在线人数突破2亿.

[34]
Avery T E, Berlin G L.Interpretation of Aerial Photographs[M]. Minneapolis: Burgess, 1977.

[35]
Anderson J R.A Land Use and Land Cover Classification System for Use with Remote Sensor Data[M]. US Government Printing Office, 1976.

[36]
肖鹏峰,刘顺喜,冯学智,等.基于中分辨率遥感图像的土地利用与覆被分类系统构建[J].中国土地科学,2006,20(2):33-38.研究目的:研制基于遥感的土地利用与覆被分类系统,为国家土地利 用与覆被基础数据库建设和宏观土地资源遥感动态监测提供分类依据.研究方法:文献资料法、比较分析法和计算机试验法.研究结果:根据中分辨率卫星传感器的 波谱响应特征和土地利用与覆被遥感信息获取技术,提出基于中分辨率遥感图像的土地利用与覆被分类系统,并给出各个类别在遥感图像上的解译标志.研究结论: 该系统基本可以满足在国家尺度上的遥感分类应用要求.

DOI

Options
文章导航

/

〈 〉