Please wait a minute...
img

官方微信

遥感技术与应用  2020, Vol. 35 Issue (5): 1089-1098    DOI: 10.11873/j.issn.1004-0323.2020.5.1089
遥感应用     
基于客观阈值与随机森林Gini指标的水体遥感指数对比
王一帆1,2(),徐涵秋1,2()
1.福州大学 环境与资源学院 空间数据挖掘与信息共享教育部重点实验室,福建 福州 350116
2.福州大学 遥感信息工程研究所 福建省水土流失遥感监测评价重点实验室,福建 福州 350116
Comparison of Remote Sensing Water Indices based on Objective Threshold Value and the Random Forest Gini Coefficient
Yifan Wang1,2(),Hanqiu Xu1,2()
1.College of Environment and Resources,Key Laboratory of Spatial Data Mining & Information Sharing of Ministry of Education,Fuzhou 350116,China
2.Institute of Remote Sensing Information Engineering,Fujian Provincial Key Laboratory of Remote Sensing of Soil Erosion,Fuzhou University,Fuzhou 350116,China
 全文: PDF(3542 KB)   HTML
摘要:

利用福建福州、西藏尼玛和澳大利亚弗伦奇3地代表不同水体类型的Sentinel-2A MSI和Landsat-8 OLI数据,采用客观阈值法(0阈值)和随机森林重要性评估法,比较和分析了改进型归一化差值水体指数(Modified Normalized Difference Water Index, MNDWI)、自动水体提取指数(Automated Water Extraction Index, AWEI)和水体指数2015 (Water Index 2015, WI2015) 这3种世界常用的水体指数之间的差异。从水体增强的效果来看,MNDWI增强的水体不仅具有丰富的信息还具有鲜明的对比度,AWEI和WI2015增强的水体信息的对比度相对偏弱。精度验证表明:3种指数提取的水体精度都较高,但MNDWI在3个地区的平均总精度略高于WI2015和AWEI,3者的平均总精度分别为91.83 %、91.16 %和90.07 %。在提取细小水体方面,MNDWI的能力强于其他2种指数,在阴影较为明显的高原山地区域,MNDWI提取水体的效果优于AWEI和WI2015。进一步采用随机森林的Gini指标进行的重要性评估表明,MNDWI在区分水体和非水体的分类中表现出了很强的重要性,尤其在Sentinel-2A MSI数据中表现得更为突出,而WI2015和AWEI的重要性则相对较弱。

关键词: 水体指数随机森林Gini指标Sentinel?2ALandsat?8评估    
Abstract:

This study used Sentinel-2A and Landsat-8 images of Fuzhou in Fujian, Nima in Tibet, China and French Island in Australia to assess the performance of three commonly-used water indices, i.e., Modified Normalized Difference Water Index (MNDWI), Automated Water Extraction Index (AWEIsh and AWEInsh) and Water Index 2015 (WI2015). The objective threshold value, i.e., 0 threshold, and random forest importance assessment method (Gini coefficient) were adopted to do the comparison with different water types (river, lake, and ocean). Among the water enhanced images of the three indices, MNDWI-enhanced water image has the highest contrast and rich information, whereas AWEI and WI2015 have relatively low contrast and are less informative. Accuracy validation shows that the water features extracted by the three indices all have high accuracy. Nevertheless, the average overall accuracy of MNDWI in the three areas is slightly higher than that of WI2015 and AWEI, which are 91.83 %, 91.16 % and 90.07 %, respectively. In addition, MNDWI can detect small water bodies and remove mountain slope shadows more effectively than the other two indices. The importance assessment revealed by the Gini coefficient of random forest further shows that MNDWI has the strongest importance in the separation of water with non-water features, especially shown in Sentinel-2A images, while WI2015 and AWEI have a relatively lower importance.

Key words: Water index    Random forest    Gini coefficient    Sentinel-2A    Landsat-8    Assessment
收稿日期: 2019-08-09 出版日期: 2020-11-26
ZTFLH:  TP79  
基金资助: 国家重点研发计划专项(2016YFA0600302)
通讯作者: 徐涵秋     E-mail: wyfan63@163.com;hxu@fzu.edu.cn
作者简介: 王一帆(1994-),男,福建福安人,硕士研究生,主要从事环境与资源遥感研究。E?mail:wyfan63@163.com
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
王一帆
徐涵秋

引用本文:

王一帆,徐涵秋. 基于客观阈值与随机森林Gini指标的水体遥感指数对比[J]. 遥感技术与应用, 2020, 35(5): 1089-1098.

Yifan Wang,Hanqiu Xu. Comparison of Remote Sensing Water Indices based on Objective Threshold Value and the Random Forest Gini Coefficient. Remote Sensing Technology and Application, 2020, 35(5): 1089-1098.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2020.5.1089        http://www.rsta.ac.cn/CN/Y2020/V35/I5/1089

图1  实验区Landsat-8 OLI 543标准假彩色影像
地区Sentinel-2A MSILandsat-8 OLI
时间轨道号时间轨道号
中国福州市2017-09-16 GMT 02:35:51205/0892017-09-16 GMT 02:32:53119/042
澳大利亚弗伦奇岛2019-01-28 GMT 00:11:110211/0732019-01-28 GMT 00:03:38092/087
中国尼玛县//2018-11-09 GMT 04:40:54140/038
表1  Sentinel-2A和Landsat-8影像数据
地区类别训练样区样本数备注
中国福州市相对浑水、非水体

S2A:水体7 046个,

非水体28 492个;

L8:水体3 256个,

非水体12 805个

同步影像
中国尼玛县相对净水、非水体

L8:水体10 121个,

非水体25 412个

澳大利亚

弗伦奇岛

相对净水、非水体

S2A:水体650个,

非水体2 280个;

L8:水体295个,

非水体1 000个

同步影像
表2  各实验区水体类型和训练样本数
图2  各实验区水体指数增强影像
图3  各实验区水体影像直方图
实验区水体指数数据最小值最大值均值标准差变异系数
福州AWEIL8-2.0330.253-0.3640.2000.549
S2A-1.3810.943-0.2890.2290.792
MNDWIL8-0.8011.000-0.3550.2820.794
S2A-0.8240.863-0.2390.2931.230
WI2015L8-119.57216.385-15.9519.2910.582
S2A-89.88660.196-13.33110.6590.800
弗伦奇岛AWEIL8-2.2230.179-0.3600.6901.917
S2A-2.3060.339-0.3860.7701.995
MNDWIL8-0.7130.8430.2470.6562.656
S2A-0.9100.9610.3160.6792.149
WI2015L8-42.3269.818-4.84516.3963.384
S2A-40.31314.522-4.29917.3844.044
尼玛县AWEI-1.2161.121-0.4080.2930.718
MNDWIL8-0.9841.000-0.2310.4501.948
WI2015-68.51762.246-20.57415.890.772
表3  各实验区不同指数影像的统计信息
精度AWEIMNDWIWI2015
福州L8漏判率/%20.4315.5916.13
误判率/%3.905.454.88
总精度/%90.5091.7991.79
Kappa0.7970.826 20.825 9
福州S2A漏判率/%6.456.453.76
误判率/%20.5516.3520.09
总精度%87.6990.0688.77
Kappa0.7510.7970.774
弗伦奇岛L8漏判率/%14.069.068.75
误判率/%1.433.003.31
总精度/%91.0193.0393.03
Kappa0.8190.858 10.857 9
弗伦奇岛S2A漏判率/%16.4111.157.74
误判率/%2.533.694.18
总精度/%89.0591.4293.07
Kappa0.7810.8260.858
尼玛县L8漏判率/%18.6316.1817.65
误判率/%2.923.3912.04
总精度/%92.1092.8389.15
Kappa0.8260.8430.766
均值漏判率/%15.2011.6910.81
误判率/%6.276.388.90
总精度/%90.0791.8391.16
Kappa0.7950.8300.816
表4  实验区各水体指数提取精度(0阈值)
图4  部分实验区各水体指数(0阈值)的局部提取结果比较(以L8为例)
图5  实验区所有特征波段的Gini指标分布排序图
特征波段福州弗伦奇岛重要性均值重要性排序
Gini值重要性Gini值重要性
AWEI469.75680.8197.57
MNDWI3318.191133.23111
WI2015945.87571.431188
Band20.55120.001312.513
Band 34.30110.001211.512
Band 40.371397.35389
Band 516.971075.26101011
Band 6264.317103.5324.54
Band 7220.35894.58466
Band 843.68981.0688.510
Band 8a1547.88490.7175.55
Band 112495.78292.9453.52
Band 121972.52390.8864.53
表5  S2A影像的Gini指标重要性排序
特征波段福州弗伦奇岛尼玛县重要性均值重要性排序
Gini值重要性Gini值重要性Gini值重要性
AWEI1 274.94257.9361 794.62445
MNDWI1 211.03372.0324 603.92121
WI20151 559.79175.7411 419.8452.332
Band 24.3180.0080.5698.338
Band 30.8390.0090.7588.679
Band 434.20756.56785.09777
Band 5243.71660.425520.9765.676
Band 6487.96460.9743 619.0823.333
Band 7376.42571.8432 432.7733.674
表6  L8影像的Gini指标重要性排序
1 Wang Meiya, Xu Hanqiu, Fu Wei, et al. Spatio-temporal Variation of Urban Surface Water and Its Influence on Urban Thermal Environment[J].Scientia Geographica Sinica, 2016, 36(7): 1099-1105.
1 王美雅, 徐涵秋, 付伟, 等. 城市地表水体时空演变及其对热环境的影响[J]. 地理科学, 2016, 36(7): 1099-1105.
2 Zhang Lei, Gong Zhaoning, Wang Qiwei, et al. Wetland Mapping of Yellow River Delta Wetlands based on Multi-feature Optimization of Sentinel-2 Images [J]. Journal of Remote Sensing, 2019, 23(2): 313-326.
2 张磊, 宫兆宁, 王启为, 等. Sentinel-2影像多特征优选的黄河三角洲湿地信息提取[J]. 遥感学报, 2019, 23(2): 313-326.
3 Zhou Y, Dong J, Xiao X, et al. Open Surface Water Mapping Algorithms: A Comparison of Water-Related Spectral Indices and Sensors [J]. Water, 2017, 9(4): 256. doi: 10.3390/w9040256.
doi: 10.3390/w9040256
4 Feyisa G L, Meilby H, Fensholt R, et al. Automated Water Extraction Index: A New Technique for Surface Water Mapping Using Landsat Imagery [J]. Remote Sensing of Environment, 2014, 140: 23-35. doi: 10.1016/j.rse.2013.08.029.
doi: 10.1016/j.rse.2013.08.029
5 Fisher A, Flood N, Danaher T. Comparing Landsat Water Index Methods for Automated Water Classification in Eastern Australia [J]. Remote Sensing of Environment, 2016, 175: 167-182. doi: 10.1016/j.rse.2015.12.055.
doi: 10.1016/j.rse.2015.12.055
6 Xu H. Modification of Normalised Difference Water Index (NDWI) to Enhance Open Water Features in Remotely Sensed Imagery [J]. International Journal of Remote Sensing, 2006,27(14):3025-3033. doi:10.1080/01431160600589179.
doi: 10.1080/01431160600589179
7 Xu Rong, Zhang Zengxiang, Zhao Chunzhe. Diffrernt Models Used in Extraction of Lake Water Body based on MODIS Data [J]. Remote Sensing information, 2015, 30(01): 111-118.
7 徐蓉, 张增祥, 赵春哲. 湖泊水体遥感提取方法比较研究[J]. 遥感信息, 2015, 30(01): 111-118.
8 Bangira T, Alfieri S M, Menenti M, et al. A Spectral Unmixing Method with Ensemble Estimation of Endmembers: Application to Flood Mapping in the Caprivi Floodplain [J]. Remote Sensing, 2017, 9(10): 1013. doi: 10.3390/rs9101013.
doi: 10.3390/rs9101013
9 Masocha M, Dube T, Makore M, et al. Surface Water Bodies Mapping in Zimbabwe Using Landsat 8 OLI Multispectral Imagery: A Comparison of Multiple Water Indices [J]. Phsics and Chemistry of the Earth, 2018, 106: 63-67. doi: 10.1016/j.pce.2018.05.005.
doi: 10.1016/j.pce.2018.05.005
10 Ryu J H, Won J S, Min K D. Waterline Extraction from Landsat TM Data in a Tidal Flat - A Case Study in Gomso Bay, Korea [J]. Remote Sensing of Environment, 2002, 83(3): 442-456. doi: 10.1016/S0034-4257(02)00059-7.
doi: 10.1016/S0034-4257(02)00059-7
11 Feng L, Hu C M, Chen X L, et al. Assessment of Inundation Changes of Poyang Lake Using MODIS Observations between 2000 and 2010 [J]. Remote Sensing of Environment, 2012, 121: 80-92. doi: 10.1016/j.rse.2012.01.014.
doi: 10.1016/j.rse.2012.01.014
12 Crist E P. A TM Tasseled Cap Equivalent Transformation for Reflectance Factor Data [J]. Remote Sensing of Environment, 1985, 17(3): 301-306. doi: 10.1016/0034-4257(85)90102-6.
doi: 10.1016/0034-4257(85)90102-6
13 McFeeters S K. The Use of the Normalized Difference Water Index (NDWI) in the Delineation of Open Water Features [J]. International Journal of Remote Sensing, 1996, 17(7): 1425-1432. doi: 10.1080/01431169608948714.
doi: 10.1080/01431169608948714
14 Danaher T, Collett L. Development, Optimisation and Multi-temporal Application of A Simple Landsat based Water Index[C]∥ Proceedings of the 13th Australasian Remote Sensing and Photogrammetry Conference, Canberra, Australia. 2006, 29.
15 Wang Dazhao, Wang Simeng, Huang Chang, et al. A Comparison of Sentinel-2 and Landsat 8 Imagery for Surface Water Extraction Using Four Common Water Indexes [J]. Remote Sensing for Land and Resources, 2019, 31(3): 1-10.
15 王大钊, 王思梦, 黄昌. Sentinel-2和Landsat-8的四种常用水体指数地表水体提取对比研究[J]. 国土资源遥感, 2019, 31(3): 1-10.
16 Murray N J, Phinn S R, DeWitt M, et al. The Global Distribution and Trajectory of Tidal Flats [J]. Nature, 2019, 565(7738): 222-236. doi: 10.1038/s41586-018-0805-8.
doi: 10.1038/s41586-018-0805-8
17 Chen J, Mueller V. Coastal Climate Change, Soil Salinity and Human Migration in Bangladesh [J]. Nature Climate Change, 2018, 8(11): 981. doi: 10.1038/s41558-018-0313-8.
doi: 10.1038/s41558-018-0313-8
18 Ogilvie A, Belaud G, Massuel S, et al. Surface Water Monitoring in Small Water Bodies: Potential and Limits of Multi-sensor Landsat Time Series [J]. Hydrology and Earth System Sciences, 2018, 22(8): 4349-4380. doi: 10.5194/hess-22-4349-2018.
doi: 10.5194/hess-22-4349-2018
19 ESA. Sentinel-2 User Handbook [EB/OL]. (2015-07-24) [2019-06-24]. , 2015.
20 ESA. Sentinel-2 MSI – Level-2A Prototype Processor Installation and User Manual [EB/OL]. (2016-04-13) [2019-06-24]. , 2016.
21 UGSG. Landsat 8 (L8) Data Users Handbook [EB/OL]. (2019-04-01) [2019-06-24]. , 2019.
22 Chavez P S. Image-based Atmospheric Corrections Revisited and Improved [J]. Photogrammetric Engineering and Remote Sensing, 1996, 62(9): 1025-1036.
23 Xu Hanqiu. Retrieval of the Reflectance and Land Surface Temperature of the Newly-launched Landsat 8 Satellite [J]. Chinese Journal of Geophysics, 2015, 58(3): 741-747.
23 徐涵秋. 新型Landsat8卫星影像的反射率和地表温度反演[J]. 地球物理学报, 2015, 58(3): 741-747.
24 Liu Hui. Regional Inequality Measurement: Methods and Evaluations [J]. Geographical Research, 2006(4): 710-718.刘慧. 区域差异测度方法与评价[J]. 地理研究, 2006(4): 710-718.
25 Tulbure M G, Broich M, Stehman S V, et al. Surface Water Extent Dynamics from Three Decades of Seasonally Continuous Landsat Time Series at Subcontinental Scale in A Semi-arid Region [J]. Remote Sensing of Environment, 2016, 178: 142-157. doi: 10.1016/j.rse.2016.02.034.
doi: 10.1016/j.rse.2016.02.034
26 Jensen J R. Introductory Digital Image Processing: A Remote Sensing Perspective[M].New Jersey:Prentice Hall Press, 2015.
27 Goldstein B A, Polley E C, Briggs F B S. Random Forests for Genetic Association Studies [J]. Statistical Applications in Genetics and Molecular Biology, 2011, 10(1). doi: 10.2202/1544-6115.1691.
doi: 10.2202/1544-6115.1691
28 Allen G H, Pavelsky T M. Global Extent of Rivers and Streams [J]. Science, 2018, 361(6402): 585-587. doi: 10.1126/science.aat0636.
doi: 10.1126/science.aat0636
[1] 李萌,年雁云,边瑞,白艳萍,马金辉. 基于多源遥感影像的青海云杉和祁连圆柏分类[J]. 遥感技术与应用, 2020, 35(4): 855-863.
[2] 张坤,刘乃文,高帅,赵书慧. 数据驱动的植被总初级生产力估算方法研究[J]. 遥感技术与应用, 2020, 35(4): 943-949.
[3] 李净,温松楠. 基于3种机器学习法的太阳辐射模拟研究[J]. 遥感技术与应用, 2020, 35(3): 615-622.
[4] 董超,赵庚星. 时序数据集构建质量对土地覆盖分类精度的影响研究[J]. 遥感技术与应用, 2020, 35(3): 558-566.
[5] 柴旭荣,李明,周义,王金风,田庆春. 影像的土地覆被快速分类[J]. 遥感技术与应用, 2020, 35(2): 315-325.
[6] 徐玉雯,张浩,陈正超,景海涛. 基于深蓝算法的Sentinel-2数据气溶胶光学厚度反演[J]. 遥感技术与应用, 2020, 35(2): 372-380.
[7] 唐廷元,付波霖,何素云,娄佩卿,闭璐. 基于GF-1和Sentinel-1A的漓江流域典型地物信息提取[J]. 遥感技术与应用, 2020, 35(2): 448-457.
[8] 肖林,车涛,戴礼云. 多源雪深数据在中国的空间特征评估[J]. 遥感技术与应用, 2019, 34(6): 1133-1145.
[9] 刘培,余志远,马威,韩瑞梅,陈正超,王涵,杨磊库. 基于地形信息的Landsat与Radarsat-2遥感数据协同分类研究[J]. 遥感技术与应用, 2019, 34(6): 1269-1275.
[10] 张帅,王振会,赵兵科,冷亮. 星载雷达在评估地基天气雷达非降水回波识别算法效果中的应用[J]. 遥感技术与应用, 2019, 34(5): 1101-1110.
[11] 黄桂平,曹艳萍. TRMM卫星3B43降水数据在黄河流域的精度分析[J]. 遥感技术与应用, 2019, 34(5): 1111-1120.
[12] 李哲,张沁雨,彭道黎. 基于高分二号遥感影像的树种分类方法[J]. 遥感技术与应用, 2019, 34(5): 970-982.
[13] 刘立, 刘勇. 基于优化分割分类层次体系的土地覆被分类制图方法探讨[J]. 遥感技术与应用, 2019, 34(1): 79-89.
[14] 谷晓天, 高小红, 马慧娟, 史飞飞, 刘雪梅, 曹晓敏. 复杂地形区土地利用/土地覆被分类机器学习方法比较研究[J]. 遥感技术与应用, 2019, 34(1): 57-67.
[15] 皋厦, 申鑫, 代劲松, 曹林. 结合LiDAR单木分割和高光谱特征提取的城市森林树种分类[J]. 遥感技术与应用, 2018, 33(6): 1073-1083.