基于独立成分分析和随机森林算法的城镇用地提取研究

doi:10.12082/dqxxkx.2020.190385

地球信息科学学报 ›› 2020, Vol. 22 ›› Issue (8): 1597-1606.doi: 10.12082/dqxxkx.2020.190385

• 地球信息科学理论与方法 • 上一篇下一篇

基于独立成分分析和随机森林算法的城镇用地提取研究

蒲东川¹^,²(), 王桂周¹^,³^,⁴, 张兆明¹^,³^,⁴^,^*(), 牛雪峰², 何国金¹^,³^,⁴, 龙腾飞¹^,³^,⁴, 尹然宇¹^,³^,⁴, 江威¹^,³^,⁴, 孙嘉悦²

1.中国科学院空天信息创新研究院, 北京 100094
2.吉林大学地球探测科学与技术学院, 长春 130026
3.海南省地球观测重点实验室, 三亚 572029
4.三亚中科遥感研究所, 三亚 572029

收稿日期:2019-07-19 修回日期:2019-11-25 出版日期:2020-08-25 发布日期:2020-10-25
通讯作者: 张兆明 E-mail:pudc17@mails.jlu.edu.cn;zhangzm@radi.ac.cn
作者简介:蒲东川（1995— ）,男,重庆奉节人,硕士生,主要从事地表覆盖分类、城市遥感等研究。 E-mail:pudc17@mails.jlu.edu.cn
基金资助:
国家自然科学基金重点项目(61731022);中科院A类先导专项(XDA19090300);国家重点研发计划课题(2016YFA0600302);国家重点研发计划课题(2016YFB0501502)

Urban Area Extraction based on Independent Component Analysis and Random Forest Algorithm

PU Dongchuan¹^,²(), WANG Guizhou¹^,³^,⁴, ZHANG Zhaoming¹^,³^,⁴^,^*(), NIU Xuefeng², HE Guojin¹^,³^,⁴, LONG Tengfei¹^,³^,⁴, YIN Ranyu¹^,³^,⁴, JIANG Wei¹^,³^,⁴, SUN Jiayue²

1. Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
2. College of Geo-exploration Science and Technology, Jilin University, Changchun 130026, China
3. Key Laboratory of Earth Observation Hainan Province, Sanya 572029, China
4. Sanya Institute of Remote Sensing, Sanya 572029, China

Received:2019-07-19 Revised:2019-11-25 Online:2020-08-25 Published:2020-10-25
Contact: ZHANG Zhaoming E-mail:pudc17@mails.jlu.edu.cn;zhangzm@radi.ac.cn
Supported by:
National Natural Science Foundation of China(61731022);Strategic Priority Research Program of the Chinese Academy of Sciences(XDA19090300);National Key Research and Development Project(2016YFA0600302);National Key Research and Development Project(2016YFB0501502)

摘要/Abstract

摘要：

城镇用地信息是联合国2030年可持续发展议程关注的重点之一。城市在世界范围内迅速扩张,快速准确地获取城镇用地信息对于政府决策具有重要作用。城镇土地覆盖信息非常复杂,包括人工建筑、树木、草地、水体等多种地表覆盖类型。基于传统人工测绘获取城镇用地信息费时费力并且难于及时更新。Landsat等遥感卫星数据为城镇用地信息提取提供了丰富的数据源。基于卫星遥感数据提取的城镇用地信息可以为未来城市的建设和管理提供基础的科学决策数据。基于监督分类方法和卫星遥感数据可快速地提取城镇用地信息,然而特征变量的选择对于高精度城镇用地信息提取尤为重要。为研究不同特征变量组合对于城镇用地信息提取的影响,以北京市为研究区,以2017年7月10日获取的Landsat 8 OLI影像为数据源,通过数据预处理、纹理提取、独立成分分析、主成分分析等得到4个维度的29个特征,选取了7种特征组合方案进行城镇用地提取。考虑随机森林算法性能稳定,分类精度高和可以方便进行特征重要性评价等优点,选择其作为监督分类算法以提取城镇用地信息,并进行了精度评定,以确定最优的城镇用地提取特征组合。研究发现：综合利用光谱特征和独立成分分析后的影像特征,提取城镇用地的总体精度为93.1%,Kappa系数为0.86,优于利用其他特征的提取结果;基于随机森林算法对数据进行训练后输出的各变量的归一化变量重要性与特征均值的标准差结果存在相似性,利用随机森林算法的变量重要性估计与特征均值折线图都可以进行变量重要性评价。

关键词: 随机森林, 独立成分分析, 主成分分析, 灰度共生矩阵, 卫星遥感, 城镇用地, Landsat 8, 特征重要性

Abstract:

Urban area information is of great significance for human development, in the 2030 United Nations (UN) Sustainable Development Agenda. Urban area expanded rapidly in many places of the world. Accurate and timely urban area information is very important for decision makers. However, land cover in urban area is highly complex, including artificial buildings, trees, grasslands, water bodies, etc. Extraction of urban land cover information based on traditional manual survey is time-consuming and difficult to update in time. Free access to remote sensing satellite data such as Landsat provides a rich source of data for urban area extraction. Urban area information extracted from space borne remote sensing images can provide basic scientific data for decision-making and city construction and management. Based on supervised classification method and satellite remote sensing data, it is possible to extract urban areas fast. However, choosing appropriate feature variables is very important for obtaining accurate urban area extraction result, especially linear correlations between different features has a significant impact on the extraction accuracy. After implementing independent component analysis (ICA) transformation to satellite remote sensing image data, linearly independent feature variables can be obtained, therefore accuracy of urban area extraction can be effectively improved. Taking Beijing city as the study area and Landsat 8 Operational Land Imager (OLI) imagery (path/row: 123/32) acquired on July 10^th, 2017 as the experimental data, preprocessing, texture extraction, independent component analysis, and principal component analysis were performed, 29 features in 4 dimensions and 7 feature variable combinations were selected. Then, Random Forest (RF) algorithm was chosen for urban area extraction owing to its stable performance, high classification accuracy and feature importance evaluation capability. Based on the random forest algorithm, feature importance evaluation, urban area extraction, and accuracy assessment were carried out to determine the optimal feature combination for urban area extraction. It was found that: (1) the overall accuracy of urban area extraction with spectral and ICA transformed features is 93.1% and the Kappa coefficient is 0.86, which is superior to the results with other features; (2) Based on the random forest algorithm, the data is trained to obtain normalized importance of each feature. There is a similarity between the normalized importance of features and the standard deviation of mean values of the features, indicating that the importance estimate of features has a close relationship with the standard deviation of mean values of the features and both can be used to estimate importance of the variables.

Key words: random forest, independent component analysis, principal component analysis, gray level co-occurrence matrix, satellite remote sensing, urban area, Landsat 8, feature importance

蒲东川, 王桂周, 张兆明, 牛雪峰, 何国金, 龙腾飞, 尹然宇, 江威, 孙嘉悦. 基于独立成分分析和随机森林算法的城镇用地提取研究[J]. 地球信息科学学报, 2020, 22(8): 1597-1606.DOI:10.12082/dqxxkx.2020.190385

PU Dongchuan, WANG Guizhou, ZHANG Zhaoming, NIU Xuefeng, HE Guojin, LONG Tengfei, YIN Ranyu, JIANG Wei, SUN Jiayue. Urban Area Extraction based on Independent Component Analysis and Random Forest Algorithm[J]. Journal of Geo-information Science, 2020, 22(8): 1597-1606.DOI:10.12082/dqxxkx.2020.190385

图/表 10

图1

图2

图3

表1

图4

图5

图6

图7

图8

表2

参考文献 26

[1]	Gong P, Liang S, Carlton E J. Urbanisation and health in China[J]. The Lancet, 2012,379(9818):843-852. doi: 10.1016/S0140-6736(11)61878-3
[2]	Weng Q. Remote sensing of impervious surfaces in the urban areas: Requirements, methods, and trends[J]. Remote Sensing of Environment, 2012,117:34-49. doi: 10.1016/j.rse.2011.02.030
[3]	徐涵秋, 王美雅. 地表不透水面信息遥感的主要方法分析[J]. 遥感学报, 2016,20(5):1270-1289.
	[ Xu H Q, Wang M Y. Remote sensing-based retrieval of ground impervious surfaces[J]. Journal of Remote Sensing, 2016,20(5):1270-1289. ]
[4]	Belgiu M, Dr Gut L. Comparing supervised and unsupervised multiresolution segmentation approaches for extracting buildings from very high resolution imagery[J]. ISPRS J Photogramm Remote Sens, 2014,96:67-75. doi: 10.1016/j.isprsjprs.2014.07.002 pmid: 25284960
[5]	Lu D, Weng Q. Use of impervious surface in urban land-use classification[J]. Remote Sensing of Environment, 2006,102(1-2):146-160. doi: 10.1016/j.rse.2006.02.010
[6]	Belgiu M, Drăguţ L. Random forest in remote sensing: A review of applications and future directions[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016,114:24-31. doi: 10.1016/j.isprsjprs.2016.01.011
[7]	Breiman L. Random forests[J]. Machine Learning, 2001,45(1):5-32. doi: 10.1023/A:1010933404324
[8]	Körting T S, Garcia Fonseca L M, Câmara G. GeoDMA-geographic data mining analyst[J]. Computers & Geosciences, 2013,57:133-145.
[9]	Belgiu M, Dr Gut L, Strobl J. Quantitative evaluation of variations in rule-based classifications of land cover in urban neighbourhoods using WorldView-2 imagery[J]. ISPRS J Photogramm Remote Sens, 2014,87(100):205-215. doi: 10.1016/j.isprsjprs.2013.11.007 pmid: 24623959
[10]	Masoumi F, Eslamkish T, Abkar A A. Integration of spectral, thermal, and textural features of ASTER data using Random Forests classification for lithological mapping[J]. Journal of African Earth Sciences, 2017,129:445-457. doi: 10.1016/j.jafrearsci.2017.01.028
[11]	Huang H, Chen Y, Clinton N. Mapping major land cover dynamics in Beijing using all Landsat images in Google Earth Engine[J]. Remote Sensing of Environment, 2017,202:166-176. doi: 10.1016/j.rse.2017.02.021
[12]	宫鹏, 张伟, 俞乐. 全球地表覆盖制图研究新范式[J]. 遥感学报, 2016,20(5):1002-1016.
	[ Gong P, Zhang W, Yu L. New research paradigm for global land cover mapping[J]. Journal of Remote Sensing, 2016,20(5):1002-1016. ]
[13]	Zhu Z, Woodcock C E, Rogan J. Assessment of spectral, polarimetric, temporal, and spatial dimensions for urban and peri-urban land cover classification using Landsat and SAR data[J]. Remote Sensing of Environment, 2012,117:72-82. doi: 10.1016/j.rse.2011.07.020
[14]	詹国旗, 杨国东, 王凤艳, 等. 基于特征空间优化的随机森林算法在GF-2影像湿地分类中的研究[J]. 地球信息科学学报, 2018,20(10):1520-1528. doi: 10.12082/dqxxkx.2018.180119
	[ Zhan G Q, Yang G D, Wang F Y, et al. The random forest classification of wetland from GF-2 imagery based on the optimized feature space[J]. Journal of Geo-information Science, 2018,20(10):1520-1528. ]
[15]	温小乐, 钟奥, 胡秀娟. 基于随机森林特征选择的城市绿化乔木树种分类[J]. 地球信息科学学报, 2018,20(12):1777-1786. doi: 10.12082/dqxxkx.2018.180310
	[ Wen X L, Zhong A, Hu X J. The classification of urban greening tree species based on feature selection of random forest[J]. Journal of Geo-information Science, 2018,20(12):1777-1786. ]
[16]	吴一全, 曹照清, 陶飞翔. 结合多尺度几何分析和KICA 的遥感图像变化检测[J]. 遥感学报, 2015,19(1):126-133.
	[ Wu Y Q, Cao Z Q, Tao F X. Change detection of remote sensing images by multi-scale geometric analysis and KICA[J]. Journal of Remote Sensing, 2015,19(1):127-133. ]
[17]	常睿春, 王璐, 王茂芝. FastICA在高光谱遥感矿物信息提取中的应用[J]. 国土资源遥感, 2013,25(4):129-132. doi: 10.6046/gtzyyg.2013.04.20
	[ Chang R C, Wang L, Wang M Z, Application of FastICA in mineral information extraction from hyperspectral remote sensing data[J]. Remote Sensing for Land and Resources, 2013,25(4):129-132.]
[18]	曹晶晶, 卓莉, 王芳. 盲信号分离技术在高光谱混合像元分解中的应用[J]. 遥感技术与应用, 2013,28(3):488-495.
	[ Cao J J, Zhuo L, Wang F. Researchon applications of blind source separation tech-niquesin hyperspectral unmixing[J]. Remote Sensing Technology and Application, 2013,28(3):488-495. ]
[19]	Hyvärinen A, Oja E. Independent component analysis: Algorithms and applications[J]. Neural Networks, 2000,13(4-5):411-430. doi: 10.1016/s0893-6080(00)00026-5 pmid: 10946390
[20]	Schneider A, Mertes C M. Expansion and growth in Chinese cities, 1978-2010[J]. Environmental Research Letters, 2014,9(2):024008. doi: 10.1088/1748-9326/9/2/024008
[21]	Gounaridis D, Chorianopoulos I, Symeonakis E. A random forest-cellular automata modelling approach to explore future land use/cover change in Attica (Greece), under different socio-economic realities and scales[J]. Sci Total Environ, 2019,646:320-335. doi: 10.1016/j.scitotenv.2018.07.302 pmid: 30055494
[22]	Li X, Gong P, Liang L. A 30-year (1984-2013) record of annual urban dynamics of Beijing City derived from Landsat data[J]. Remote Sensing of Environment, 2015,166:78-90. doi: 10.1016/j.rse.2015.06.007
[23]	Paul A, Mukherjee D P, Das P. Improved random forest for classification[J]. IEEE Trans Image Process, 2018,15(6):36-45.
[24]	Qihao W, Xuefei H. Medium spatial resolution satellite imagery for estimating and mapping urban impervious surfaces using LSMA and ANN[J]. IEEE Transactions on Geoscience and Remote Sensing, 2008,46(8):2397-2406. doi: 10.1109/TGRS.2008.917601
[25]	Teluguntla P. A 30-m landsat-derived cropland extent product of Australia and China using random forest machine learning algorithm on Google Earth Engine cloud computing platform[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2018,17(3):273-289.
[26]	Xu Z, Chen J, Xia J. Multisource earth observation data for land-cover classification using random forest[J]. IEEE Geoscience and Remote Sensing Letters, 2018,15(5):789-793. doi: 10.1109/LGRS.2018.2806223

序号		特征名称
a	光谱维度	Coastal、Blue、Green、Red、NIR、SWIR1、SWIR2
b	ICA维度	ICA1、ICA2、ICA3、ICA4、ICA5、ICA6、ICA7
c	PCA维度	PCA1、PCA2、PCA3、PCA4、PCA5、PCA6、PCA7
d	纹理维度	Mean.(Mean)、Var.(Variance)、Hom.(Homogeneity)、Con.(Contrast)、Dis.(Dissimilarity)、Ent.(Entropy)、ASM.(Angular Second Moment)、Cor.(Correlation)

序号	特征数量	特征名称	总体精度/%	Kappa系数	错分率/%	漏分率/%	运行时间/s
1	7	光谱特征	84.3	0.71	29.2	19.6	101.1
2	8	纹理特征	39.5	0.27	75.6	33.3	132.3
3	7	PCA特征	81.9	0.62	25.3	23.9	97.5
4	7	ICA特征	87.1	0.78	20.8	20.2	103.7
5	15	光谱+纹理	86.3	0.69	30.5	28.3	197.8
6	14	光谱+PCA	89.2	0.79	23.3	31.5	164.2
7	14	光谱+ICA	93.1	0.86	17.5	18.4	157.1

基于独立成分分析和随机森林算法的城镇用地提取研究

Urban Area Extraction based on Independent Component Analysis and Random Forest Algorithm

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 26

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	赵鹏军, 曹毓书. 基于多源地理大数据与机器学习的地铁乘客出行目的识别方法[J]. 地球信息科学学报, 2020, 22(9): 1753-1765.
[2]	朱守杰, 杜世宏, 李军, 商硕硕, 杜守基. 融合多源空间数据的城镇人口分布估算[J]. 地球信息科学学报, 2020, 22(8): 1607-1616.
[3]	李玉, 李奕燃, 王光辉, 石雪. 基于加权指数函数模型的高光谱图像分类方法[J]. 地球信息科学学报, 2020, 22(8): 1642-1653.
[4]	李婉, 牛陆, 陈虹, 吴骅. 基于随机森林算法的地表温度鲁棒降尺度方法[J]. 地球信息科学学报, 2020, 22(8): 1666-1678.
[5]	毛亚萍, 房世峰. 基于机器学习的参考作物蒸散量估算研究[J]. 地球信息科学学报, 2020, 22(8): 1692-1701.
[6]	崔成, 任红艳, 赵璐, 庄大方. 基于街景影像多特征融合的广州市越秀区街道空间品质评估[J]. 地球信息科学学报, 2020, 22(6): 1330-1338.
[7]	杨建思, 柳帅, 王艳东, 廖明生. 融合多源大数据的武汉城市空置区域评估与分析[J]. 地球信息科学学报, 2020, 22(5): 997-1007.
[8]	曹泽涛, 方子东, 姚瑾, 熊礼阳. 基于随机森林的黄土地貌分类研究[J]. 地球信息科学学报, 2020, 22(3): 452-463.
[9]	黄康, 戴文远, 黄万里, 欧惠. 基于FLUS模型与动能定理的城镇用地增长边界划定[J]. 地球信息科学学报, 2020, 22(3): 557-567.
[10]	罗耀文, 任周鹏, 葛咏, 韩李涛, 刘梦晓, 何亚文. 基于PCA-GWR方法的村级贫困时空格局及致贫因素分析[J]. 地球信息科学学报, 2020, 22(2): 231-245.
[11]	郭云开, 张晓炯, 许敏, 刘雨玲, 钱佳, 章琼. 路域植被等效水厚度估算模型研究[J]. 地球信息科学学报, 2020, 22(2): 308-315.
[12]	周佳, 赵亚鹏, 岳天祥, 卢涛. 结合HASM和GWR方法的省级尺度近地表气温估算[J]. 地球信息科学学报, 2020, 22(10): 2098-2107.
[13]	陈炜, 黄慧萍, 田亦陈, 杜云艳. 基于Google Earth Engine平台的三江源地区生态环境质量动态监测与分析[J]. 地球信息科学学报, 2019, 21(9): 1382-1391.
[14]	王法溧,何晓宇,方泽兴,魏显虎,叶虎平,施令飞. 基于长时间序列遥感数据的阿伯德尔国家公园生态环境质量评价[J]. 地球信息科学学报, 2019, 21(9): 1479-1479.
[15]	耿仁方,付波霖,蔡江涛,陈晓雨,蓝斐芜,余杭洺,李青逊. 基于无人机影像和面向对象随机森林算法的岩溶湿地植被识别方法研究[J]. 地球信息科学学报, 2019, 21(8): 1295-1306.