基于三维计盒法的城市空间形态分维计算和分析
作者简介:秦静(1986- ),女,山东济南人,博士研究生,主要研究方向为城市与区域规划,城市空间增长测度与模拟。E-mail:qjing1986@163.com
收稿日期: 2014-05-03
要求修回日期: 2014-10-11
网络出版日期: 2015-01-10
基金资助
国家自然科学基金项目(41401164)
国家“十二五”科技支撑计划项目(2012BAJ22B03-04)
A three dimensional box-counting method for estimating fractal dimension of urban form
Received date: 2014-05-03
Request revised date: 2014-10-11
Online published: 2015-01-10
Copyright
将城市形态分形维数的二维盒维数方法扩展到三维空间,提出了一种三维盒维数计算方法;并且针对以往城市分形研究中人工判定无尺度区不精确的问题,借鉴了二阶导数的方法进行分形无尺度区的自动识别。以扬州市中心城区2003年与2012年两期数据为例,详细介绍了城市三维空间形态分维计盒法的计算过程。实验结果显示,通过二阶导数方法能够有效地自动识别无尺度区;2003年与2012年数据无尺度区线性拟合确定系数R2均在0.996以上,扬州市三维空间形态具有明确的分形特征;2012年较2003年三维分维值的增加表明扬州市城市三维空间利用总体上趋于有效与紧凑。
秦静 , 方创琳 , 王洋 , 李秋颖 , 张永姣 . 基于三维计盒法的城市空间形态分维计算和分析[J]. 地理研究, 2015 , 34(1) : 85 -96 . DOI: 10.11821/dlyj201501008
Urban fractal dimension analysis is an important method for quantifying the measurement of urban morphology. With rapid urban development in China in recent two decades, the three-dimensional characteristic has been the essential feature of urban morphology. However the research of urban fractal dimension are mostly focused on the two-dimensional space, few studies have been conducted from a 3D perspective. In this paper, a three-dimensional box-counting method is proposed to estimate fractal dimension of urban form. The scale-free region is usually determined by experience in the classical urban fractal research. In the lnr-lnN(r) curve, the researchers find out a line segment with good linear relationship over a range of length scales with their experience as the scaling range. This identification method is subjective, and cannot accurately identify and calculate the fractal dimension. Moreover, for the same set of data, different scale ranges could be attained by different observers. In this paper, a method based on the second-order derivative of log-log curves is used to automatically identify the fractal scale-free range. The central urban area of Yangzhou city is selected as the case area. Two high resolution remote sensing images of 2003 and 2012, which are respectively the QuickBird data with 0.61 m resolution and the GeoEye data with 0.5 m resolution, are used to acquire Yangzhou three-dimensional morphology data. Then based on the data, the detailed process of the three-dimensional box-counting method is described. The experimental results show that: the scaling range is the basis of calculating city fractals, the second-order derivative method can automatically identify fractal scale-free range accurately; The linear regression coefficients of scale-free range R2 are all over 0.996 in the two years, which demonstrates that Yangzhou three-dimensional morphology is fractal; The fractal dimension of urban morphology has increased significantly in 2012 compared with the year 2003, which indicates that the utilization of 3D spaces has become more efficient and compact; The three-dimensional box-counting algorithm proposed in this paper not only provides a new methodology in quantifying the measurement of urban spatial structure and the evolution of urban morphology, but also extends the scope of urban studies to higher dimensions.
Fig. 1 The study area in Yangzhou city图1 扬州中心城区研究范围 |
Fig. 2 Yangzhou 3D morphology data acquisition图2 扬州市三维数据获取流程图 |
Fig. 3 The 3D model of different types of buildings in Yangzhou图3 扬州市2003年与2012年不同类型建筑三维模型 |
Tab. 1 Data of N(r), length and height of the box of different types of buildings in Yangzhou in 2003表1 2003年扬州市不同类型建筑各尺度下非空盒子数N(r)、盒子边长与高度数据 |
| 数据编号 | 栅格尺度r | 盒子边长(m) | 盒子高度(m) | 非空盒子数N(r) | |||
|---|---|---|---|---|---|---|---|
| 全部建筑 | 住宅建筑 | 商业及公共建筑 | 工业建筑 | ||||
| 1 | 32768 | 16820.0 | 84/1 | 1 | 1 | 1 | 1 |
| 2 | 16384 | 8410.0 | 84/2 | 8 | 7 | 7 | 4 |
| 3 | 8192 | 4205.0 | 84/22 | 27 | 22 | 24 | 14 |
| 4 | 4096 | 2102.5 | 84/23 | 114 | 81 | 97 | 52 |
| 5 | 2048 | 1051.3 | 84/24 | 532 | 378 | 403 | 184 |
| 6 | 1024 | 525.6 | 84/25 | 2506 | 1713 | 1734 | 677 |
| 7 | 512 | 262.8 | 84/26 | 12669 | 8550 | 7386 | 2652 |
| 8 | 256 | 131.4 | 84/27 | 67630 | 43460 | 30830 | 11104 |
| 9 | 128 | 65.7 | 84/28 | 374528 | 234004 | 129130 | 53437 |
| 10 | 64 | 32.9 | 84/29 | 2134914 | 1351850 | 596336 | 272262 |
| 11 | 32 | 16.4 | 84/210 | 12766627 | 8339106 | 2979241 | 1535577 |
| 12 | 16 | 8.2 | 84/211 | 75937136 | 49700451 | 16812733 | 9312826 |
| 13 | 8 | 4.1 | 84/212 | 477929911 | 309214618 | 106189849 | 61164739 |
| 14 | 4 | 2.1 | 84/213 | 3304628089 | 2115804512 | 741251507 | 437106020 |
| 15 | 2 | 1.0 | 84/214 | 24422837158 | 15537396886 | 5513774766 | 3291625360 |
| 16 | 1 | 0.5 | 84/215 | 187453250141 | 118859270858 | 42462081155 | 25510807078 |
Tab. 2 Data of N(r), length and height of the box of different types of buildings in Yangzhou in 2012表2 2012年扬州市不同类型建筑各尺度下非空盒子数N(r)、盒子边长与高度数据 |
| 数据编号 | 栅格尺度r | 盒子边长(m) | 盒子高度(m) | 非空盒子数N(r) | |||
|---|---|---|---|---|---|---|---|
| 全部建筑 | 住宅建筑 | 商业及公共建筑 | 工业建筑 | ||||
| 1 | 32768 | 16820.0 | 84/1 | 1 | 1 | 1 | 1 |
| 2 | 16384 | 8410.0 | 84/2 | 8 | 8 | 8 | 4 |
| 3 | 8192 | 4205.0 | 84/22 | 40 | 35 | 34 | 18 |
| 4 | 4096 | 2102.5 | 84/23 | 209 | 165 | 148 | 81 |
| 5 | 2048 | 1051.3 | 84/24 | 1060 | 851 | 637 | 317 |
| 6 | 1024 | 525.6 | 84/25 | 5410 | 4181 | 2661 | 1275 |
| 7 | 512 | 262.8 | 84/26 | 28203 | 20628 | 11316 | 5387 |
| 8 | 256 | 131.4 | 84/27 | 152759 | 106897 | 48168 | 25202 |
| 9 | 128 | 65.7 | 84/28 | 862511 | 592261 | 206508 | 128411 |
| 10 | 64 | 32.9 | 84/29 | 5011672 | 3463201 | 991332 | 682157 |
| 11 | 32 | 16.4 | 84/210 | 29344515 | 20398459 | 5081260 | 3960058 |
| 12 | 16 | 8.2 | 84/211 | 171021820 | 116916801 | 29023973 | 24695065 |
| 13 | 8 | 4.1 | 84/212 | 1072801626 | 718327115 | 184645076 | 166275801 |
| 14 | 4 | 2.1 | 84/213 | 7419541979 | 4894512027 | 1293309246 | 1204865166 |
| 15 | 2 | 1.0 | 84/214 | 54842418139 | 35857746230 | 9633949384 | 9145214151 |
| 16 | 1 | 0.5 | 84/215 | 420958538209 | 273922432973 | 74261157130 | 71175761698 |
Tab. 3 Evolution of the fractal dimension of different types of buildings in Yangzhou表3 扬州市2003年与2012年不同类型建筑分维计算结果 |
| 范围 | 参数 | 全部建筑 | 居住建筑 | 商业与公共建筑 | 工业建筑 | |
|---|---|---|---|---|---|---|
| 2003年 | 全部序列 | D | 2.4573 | 2.4244 | 2.2798 | 2.2740 |
| R2 | 0.9974 | 0.9965 | 0.9956 | 0.9920 | ||
| 无尺度区 | D | 2.4161 | 2.3375 | 2.1111 | 2.0389 | |
| R2 | 0.9988 | 0.9989 | 0.9990 | 0.9968 | ||
| 2012年 | 全部序列 | D | 2.5218 | 2.4827 | 2.3196 | 2.3745 |
| R2 | 0.9990 | 0.9986 | 0.9963 | 0.9951 | ||
| 无尺度区 | D | 2.4624 | 2.4279 | 2.1621 | 2.1505 | |
| R2 | 0.9996 | 0.9994 | 0.9991 | 0.9985 |
Fig. 4 The fractal dimension of Yangzhou three-dimensional morphology in 2003图4 2003年扬州城市形态三维分维分析图 |
Fig. 5 The fractal dimension of Yangzhou three-dimensional morphology in 2012图5 2012年扬州城市形态三维分维分析图 |
The authors have declared that no competing interests exist.
| [1] |
|
| [2] |
|
| [3] |
|
| [4] |
[
|
| [5] |
[
|
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
[
|
| [10] |
[
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
[
|
| [15] |
[
|
| [16] |
|
| [17] |
[
|
| [18] |
|
| [19] |
|
| [20] |
[
|
| [21] |
|
| [22] |
[
|
| [23] |
[
|
| [24] |
[
|
| [25] |
[
|
/
| 〈 |
|
〉 |