Please wait a minute...
img

官方微信

遥感技术与应用  2020, Vol. 35 Issue (6): 1292-1302    DOI: 10.11873/j.issn.1004-0323.2020.6.1292
冰雪遥感专栏     
基于特征优选的GF-3全极化数据积雪识别
马腾耀(),肖鹏峰(),张学良,马威,郭金金
南京大学地理与海洋科学学院,自然资源部国土卫星遥感应用重点实验室,江苏省地理信息技术重点实验室,江苏 南京 210023
Recognition of Snow Cover based on Features Selectionin GF-3 Fully Polarimetric Data
Tengyao Ma(),Pengfeng Xiao(),Xueliang Zhang,Wei Ma,Jinjin Guo
Jiangsu Provincial Key Laboratory of Geographic Information Science and Technology,Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources,School of Geography and Ocean Science,Nanjing University,Nanjing 210023,China
 全文: PDF(15272 KB)   HTML
摘要:

以新疆阿尔泰山南麓克兰河流域典型区为研究区,利用GF-3全极化数据进行积雪探测,提出了一种基于特征优选的积雪识别方法。首先通过极化分解获取了GF-3数据的22个极化特征,并利用随机森林方法计算各特征的重要性,构建特征优选规则生成最优特征集,然后基于最优特征集对积雪进行识别。分析特征的重要性发现,同极化后向散射系数对积雪识别的贡献比交叉极化的贡献大,面散射和体散射对积雪识别的贡献比二面角散射贡献大。将该方法与最大似然法、支持向量机、BP神经网络3种分类器的对比发现,使用最优特征集并且利用随机森林方法的积雪识别精度最高(F指数为0.86,总体精度为0.79)。结果表明:基于特征优选进行积雪识别,不仅使得积雪识别效率得到提高,而且保持精度不变甚至有所增加,证明了该方法在积雪识别中的有效性。

关键词: 积雪识别GF?3极化分解特征优选随机森林    
Abstract:

This study proposed a recognition method for snow cover based on feature selection using GF-3 fully polarimetric data. The study area was selected from the typical area of the Kelan River Basin in the southern piedmont of the Altai Mountains, Xinjiang Province. First, we obtained 22 polarization features of GF-3 data by polarization decomposition. The importance of each feature was calculated by using Random Forest (RF) method. Then, we designed the rules of feature selection to generate the optimal feature sets, which were used to recognize snow cover with RF method. Analyzing the importance of the features, we can find that, for snow cover recognition, the contribution of the same polarization backscattering coefficient is greater than that of the cross polarization backscattering coefficient, and the contribution of the surface scattering or volume scattering is greater than that of the dihedral angle scattering. Finally, a comparison with the Maximum Likelihood, Support Vector Machine, and BP neural network was made for testing the performance of the proposed method. It is found that the optimal feature sets using RF method to recognize snow cover have the highest accuracy (F-score is 0.86, overall accuracy is 0.79). From the selection of classifiers and the results of features selection, the proposed method is very effective in recognition of snow cover.

Key words: Snow cover recognition    GF-3    Polarization decomposition    Features selection    Random Forest
收稿日期: 2019-12-27 出版日期: 2021-01-26
ZTFLH:  TP181  
基金资助: 国家自然科学基金项目(41671344);国家科技基础资源调查专项课题(2017FY100502)
通讯作者: 肖鹏峰     E-mail: tengyaom@gmail.com;xiaopf@nju.edu.cn
作者简介: 马腾耀(1995-),男,山西长治人,硕士研究生,主要从事积雪遥感和数字图像处理研究。E?mail:tengyaom@gmail.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  
马腾耀
肖鹏峰
张学良
马威
郭金金

引用本文:

马腾耀,肖鹏峰,张学良,马威,郭金金. 基于特征优选的GF-3全极化数据积雪识别[J]. 遥感技术与应用, 2020, 35(6): 1292-1302.

Tengyao Ma,Pengfeng Xiao,Xueliang Zhang,Wei Ma,Jinjin Guo. Recognition of Snow Cover based on Features Selectionin GF-3 Fully Polarimetric Data. Remote Sensing Technology and Application, 2020, 35(6): 1292-1302.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2020.6.1292        http://www.rsta.ac.cn/CN/Y2020/V35/I6/1292

图1  研究区位置图
图2  训练样本和验证样本在光学遥感图像上的分布
图3  不同下垫面类型积雪、无积雪覆盖地表后向散射系数分布
类型特征
后向散射系数HHHVVHVV
Pauli分解PoddPdblPvol
H-A-αˉ分解HAαˉλˉ
Freeman分解FdblFvolFodd
Yamaguchi分解YdblYvolYoddYhlx
Anyang分解AdblAvolAoddAhlx
表1  特征列表
图4  特征优选前的积雪识别结果
分类器准确率(P)召回率(R)F指数(F)总体精度(A)
随机森林0.870.840.860.79
最大似然法0.880.770.830.75
支持向量机0.900.740.810.74
BP神经网络0.820.660.730.63
表2  特征优选前不同分类器的识别精度比较
图5  特征重要性排序(H3表示H-A-αˉ极化分解的第3个分量,P3表示Pauli分解的第3个分量,其他缩写类似)
图6  特征个数及积雪识别总体精度图
  图7特征优选后的积雪识别结果
图8  Landsat 8和GF-3图像及积雪识别结果细节图
分类器准确率(P)召回率(R)F指数(F)总体精度(A)
随机森林0.850.870.860.79
最大似然法0.880.790.830.76
支持向量机0.890.770.820.75
BP神经网络0.870.760.810.73
表3  特征优选后不同分类器的识别精度比较
1 Cai Dihua,Guo Ni,Wang Xing,et al. The Spatial and Temporal Variations of Snow Cover over the Qilian Mountains based on MODIS Data[J]. Journal of Glaciology and Geocryology,2009,31(6):1028-1036.
1 蔡迪花,郭妮,王兴,等. 基于MODIS的祁连山区积雪时空变化特征[J]. 冰川冻土,2009,31(6):1028-1036.
2 Wang Jian. Comparison and Analysis on Methods of Snow Cover Mapping by Using Satellite Remote Sensing Data[J]. Remote Sensing Technology and Application,1999,14(4):29-36.
2 王建. 卫星遥感雪盖制图方法对比与分析[J]. 遥感技术与应用,1999,14(4):29-36.
3 Cui Caixia,Wei Rongqing,Li Yang. Long-term Change of Seasonal Snowcover and Its Effects on Runoff Volume in the Upper Reaches of the Tarim River[J]. Arid Land Geography,2005,28(5):569-573.
3 崔彩霞,魏荣庆,李杨. 塔里木河上游地区积雪长期变化趋势及其对径流量的影响[J]. 干旱区地理,2005,28(5):569-573.
4 Wang Shijin,Ren Jiawen. A Review of the Progresses of Avalanche Hazards Research[J]. Progress in Geography,2012,31(11):1529-1536.
4 王世金,任贾文. 国内外雪崩灾害研究综述[J]. 地理科学进展,2012,31(11):1529-1536.
5 Lou Mengyun,Liu Zhihong,Lou Shaoming,et al. Temporal and Spatial Distribution of Snow Cover in Xinjiang from 2002 to 2011[J]. Journal of Glaciology and Geocryology,2013,35(5):1095-1102.
5 娄梦筠,刘志红,娄少明,等. 2002~2011年新疆积雪时空分布特征研究[J]. 冰川冻土,2013,35(5):1095-1102.
6 Nagler T,Rott H. Retrieval of Wet Snow by Means of Multitemporal SAR Data[J]. IEEE Transactions on Geoscience and Remote Sensing,2000,38(2):754-765.
7 Xiao Pengfeng,Feng Xuezhi,Xie Shunping,et al. Research Progresses of High-resolution Remote Sensing of Snow in Manasi River Basin in Tianshan Mountains,Xinjiang Province[J]. Journal of Nanjing University (Natural Sciences Edition),2015,51(5):909-920.
7 肖鹏峰,冯学智,谢顺平,等. 新疆天山玛纳斯河流域高分辨率积雪遥感研究进展[J]. 南京大学学报(自然科学版),2015,51(5):909-920.
8 Huang L,Li Z,Tian B S,et al. Recognition of Supraglacial Debris in the Tianshan Mountains on Polarimetric SAR Images[J].Remote Sensing of Environment, 2014,145(8):47-54.
9 Schellenberger T,Ventura B,Zebisch M,et al. Wet Snow Cover Mapping Algorithm based on Multitemporal COSMO-SkyMed X-Band SAR Images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2012,5(3):1-9.
10 Rott H,Davis R E,Multifrequency and Polarimetric SAR Observations on Alpine Glaciers[J]. Annals of Glaciology,1993,17:98-104.
11 Li Zhen,Guo Huadong,Li Xinwu,et al. SAR Interferometry Coherence Analysis and Snow Mapping[J]. Journal of Remote Sensing,2002,6(5):334-338.
11 李震,郭华东,李新武,等. SAR干涉测量的相干性特征分析及积雪划分[J]. 遥感学报,2002,6(5):334-338.
12 Singh G,Venkataraman G. Application of Incoherent Target Decomposition Theorems to Classify Snow Cover over the Himalayan Region[J]. International Journal of Remote Sensing,2012,33(13):4161-4177.
13 Huang L,Li Z,Tian B S,et al. Classification and Snow Line Detection for Glacial Areas Using the Polarimetric SAR Image[J]. Remote Sensing of Environment,2011,115(7):1721-1732.
14 Tsai Y S,Dietz A,Oppelt N,et al. Wet and Dry Snow Detection Using Sentinel-1 SAR Data for Mountainous Areas with a Machine Learning Technique[J]. Remote Sensing,2019,11(8):895. doi:10.3390/rs11080895.
doi: 10.3390/rs11080895
15 Pal M. Random Forest Classifier for Remote Sensing Classification[J]. International Journal of Remote Sensing,2005,26(1):217-222.
16 Chen J,Chen J,Liao A P,et al. Global Land Cover Mapping at 30 m Resolution: A POK-based Operational Approach[J]. ISPRS Journal of Photogrammetry and Remote Sensing,2015,103:7-27.
17 Shen Yongping,Wang Guoya,Su Hongchao,et al. Hydrological Processes Responding to Climate Warming in the Upper Reaches of Kelan River Basin with Snow-dominated of the Altay Mountains Region,Xinjiang,China[J]. Journal of Glaciology and Geocryology,2007,29(6):845-854.
17 沈永平,王国亚,苏宏超,等. 新疆阿尔泰山区克兰河上游水文过程对气候变暖的响应[J]. 冰川冻土,2007,29(6):845-854.
18 Zhou Bocheng. An Analysis on the Relationship between Streamflow and Precipitation in Altay Mountains Region[J]. Journal of Glaciology and Cryopenology,1983,5(4):49-56.
18 周伯诚. 我国阿尔泰山的降水及河流径流分析[J]. 冰川冻土,1983,5(4):49-56.
19 He Guangjun,Feng Xuezhi,Xiao Pengfeng,et al. Characterization of C band SAR Image for Snow in Mountainous Areas of Manasi River Basin[J]. Journal of Nanjing University (Natural Sciences Edition),2015,51(5):955-965.
19 玛纳斯河流域山区积雪的C波段SAR图像表征[J]. 南京大学学报(自然科学),2015,51(5):955-965.
20 Wang Chao,Zhang Hong,Chen Xi,et al. Image Processing of Fully Polarimetric SAR[M]. Beijing: Science Press,2008.王超,张红,陈曦,等. 全极化合成孔径雷达图像处理[M]. 北京: 科学出版社,2008.
21 Cloude S R,Pottier E. An Entropy based Classification Scheme for Land Application of Polarimetric SAR[J]. IEEE Transactions on Geoscience and Remote Sensing,1997,35(1):68-78.
22 Freeman A,Durden S L. A Three-Component Scattering Model for Polarimetric SAR Data[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998,36(3):963-973.
23 Yamaguchi Y,Moriyama T,Ishido M,et al. Four-Component Scattering Model for Polarimetric SAR Image Decomposition[J]. IEEE Transactions on Geoscience and Remote Sensing,2005,104(8):1699-1706.
24 Yajima Y,Yamaguchi Y,Yamada H,et al. A Four-Component Decomposition of POLSAR Images based on the Coherency Matrix[J]. IEEE Geoscience and Remote Sensing Letters,2006,3(3):292-296.
25 An W,Yi C,Jian Y. Three-Component Model-based Decomposition for Polarimetric SAR Data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2010,48(6):2732-2739.
26 An W,Xie C,Yuan X,et al. Four-Component Decomposition of Polarimetric SAR Images with Deorientation[J].IEEE Geoscience and Remote Sensing Letters,2011,8(6):1090-1094.
27 Breiman L. Random Forests[J]. Machine Learning,2001,45(1):5-32.
28 Xu Qiao,Zhang Xiao,Yu Shaohuai,et al. Multi-feature-based Classification Method Using Random Forest and Superpixels for Polarimetric SAR Images[J]. Journal of Remote Sensing,2019,23(4):685-694.
28 徐乔,张霄,余绍淮,等. 综合多特征的极化SAR图像随机森林分类算法[J]. 遥感学报,2019, 23(4):685-694.
29 Du P J,Samat A,Waske B,et al. Random Forest and Rotation Forest for Fully Polarized SAR Image Classification Using Polarimetric and Spatial Features[J]. ISPRS Journal of Photogrammetry and Remote Sensing,2015,105:38-53. doi: 101016/j.is prsjprs.2015.03.002.
doi: 101016/j.is prsjprs.2015.03.002
30 Tang Tingyuan,Fu Bolin,He Suyun,et al. Identification of Typical Land Features in the Lijiang River Basin with Fusion Optics and Radar[J]. Remote Sensing Technology and Application,2020,35(2):448-457.
30 唐廷元,付波霖,何素云,等. 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.
31 Cherkassky V,Ma Y. Practical Selection of SVM Parameters and Noise Estimation for SVM Regression[J]. Neural Networks,2004,17(1):113-126.
[1] 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.
[2] 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.
[3] 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.
[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] 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.
[6] 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.
[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] 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.
[9] 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.
[10] 李春江,沈国状,张继超. 基于灰色系统理论的植被物理参数与极化分解参数的关联分析[J]. 遥感技术与应用, 2019, 34(4): 839-846.
[11] 李春江, 沈国状, 张继超. 基于灰色系统理论的植被物理参数与极化分解参数的关联分析—以鄱阳湖湿地为例[J]. 遥感技术与应用, 2019, 34(2): 284-292.
[12] 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.
[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] 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.
[15] 陈伟民,张凌,宋冬梅,王斌,丁亚雄,许明明,崔建勇. 基于AdaBoost改进随机森林的高光谱图像地物分类方法研究[J]. 遥感技术与应用, 2018, 33(4): 612-620.