一种基于集成学习和特征融合的遥感影像分类新方法

针对多源遥感数据分类的需要，提出了一种基于全极化SAR影像、极化相干矩阵特征、光学遥感影像光谱和纹理的多种特征融合和多分类器集成的遥感影像分类新方法.对全极化PALSAR数据进行预处理和极化相干矩阵特征提取，利用灰度共生矩阵计算光学和SAR影像的对比度、逆差距、二阶距、差异性等纹理特征参数，并与光谱特征结合，形成6种组合策略.利用集成学习方法对随机森林分类器、子空间分类器、最小距离分类器、支持向量机分类器、反向传播神经网络分类器等分类器进行组合，对不同组合策略的遥感影像特征集进行分类.结果表明提出的基于多种特征和多分类器集成的新方法很好地利用了主被动遥感数据在不同地表景观类型提取上的潜力，综合了多种算法的优势，能够有效地提高总体精度和各类别的分类精度.

Abstract

To make full use of the multi-source remotely sensed data for classification, a novel method was proposed based on the integration of full-polarization SAR (HH, HV, VH, VV) data, features of polarization coherence matrix, spectral features provided by optical data, texture features extracted from optical and SAR data and multi-classifier ensemble. Preprocessing for full-polarization data was performed and polarimetric features are extracted from polarization coherence matrix. Spatial textural features including contrast, dissimilarity, second moment, etc., are extracted from PALSAR full-polarization data and optical image using Grey-level Co-occurrence Matrix (GLCM) method. Features of polarization coherency matrix, full-polarization SAR channels, spectral and textures are integrated by 6 strategies. Some well-known classification techniques, including Support Vector Machine (SVM), Minimum Distance (MD), Back Propagation Neural Network (BPNN), Multi-Layer Perceptron (MLP), Random Subspace (RSS), Random Forest (RF) classifiers were selected to test different combination strategies. The parallel and sequential ensemble learning techniques were selected to integrate single classifier for land cover classification. The results indicate that the proposed approach integrating multi-source, multi-features and multi-classifier strategy can make full use of the potential of optical and SAR remotely sensed data for landscape types, and improve the overall accuracy and the accuracy of single land cover type effectively.

参考文献

[1] Du Peijun, Liu Pei, Gamba P. Urban spatial and temporal changes analysis based on spectral, polarimetric, temporal, spatial dimensions and decision level fusion［C］.1st EARSel Workshop on temporal analysis of satellite images, Mykonos, Greece, 2225 May 2012, 5864.

[2] Yu Fan, Li Haitao, Wan Zi. Synthesis of multi-source remote sensing data for classification based on bayesian theory and Mrf［J］. Journal of remtoe sensing, (余凡, 李海涛，万紫.结合贝叶斯理论和MRF的主被动遥感数据协同分类.遥感学报) 2012,16(4)： 809826.

[3] Brunner D, Lemoine G, Bruzzone L, et al.Building Height Retrieval From VHR SAR Imagery Based on an Iterative Simulation and Matching Technique［J］. Geoscience and Remote Sensing, IEEE Transactions on, 2010,48(3)： 14871504.

[4] Zhong Ping, Wang Runsheng. A Multiple Conditional Random Fields Ensemble Model for Urban Area Detection in Remote Sensing Optical Images［J］. Geoscience and Remote Sensing, IEEE Transactions on, 2007,45(12)： 39783988.

[5] Lehmann A, Caccetta A, Zhou Z, et al. Joint Processing of Landsat and ALOS-PALSAR Data for Forest Mapping and Monitoring［J］. Geoscience and Remote Sensing, IEEE Transactions on, 2012,50(1)： 5567.

[6] Sportouche H, Tupin F, Denise L. Building detection by fusion of optical and SAR features in metric resolution data［C］, Geoscience and Remote Sensing Symposium, Cape Town, 1217 July 2009, IV769IV772.

[7] Brunner D, Lemoine G, Bruzzone L. Earthquake Damage Assessment of Buildings Using VHR Optical and SAR Imagery［J］. Geoscience and Remote Sensing. IEEE Transactions, 2010,48(5)： 24032420.

[8] Gamba P, DellAcqua F, Dasarathy V. Urban remote sensing using multiple data sets： Past, present, and future［J］. Information Fusion, 2005,6(4)： 319326.

[9] Dong Jiang, Zhuang Dafang, Huang Yaohuan et al. Advances in Multi-Sensor Data Fusion： Algorithms and Applications［J］. Sensors, 2009,9(10)： 77717784.

[10] Zhang Hongsheng, Zhang Yuanzhi, Lin Hui. Urban land cover mapping using random forest combined with optical and SAR data［C］. Geoscience and Remote Sensing Symposium (IGARSS), Munich, 2227 July 2012, 68096812.

[11] Rusmini M, Candiani G, Frassy F, et al. High-resolution SAR and high-resolution optical data integration for sub-urban land-cover classification［C］. Geoscience and Remote Sensing Symposium, Munich, 2227, July, 2012,49864989.

[12] Gamba P, Liu P, Du P, et al. Evaluation and analysis of fusion algorithms for active and passive remote sensing image［C］. Geoscience and Remote Sensing Symposium, Munich, 2227, July, 2012, 22722275.

[13] Rodriguez-Galiano F, Chica-Olmo M, Abarca-Hernandez F, et al. Random Forest classification of Mediterranean land cover using multi-seasonal imagery and multi-seasonal texture［J］. Remote Sensing of Environment, 2012, (121)： 93107.

[14] Jiao Longhao, Zhou Zhongfa, Li Bo. Study of SAR Image Texture Feature Extraction Based on GLCM in Guizhou Karst Mountainous Region.in Remote Sensing［C］, Environment and Transportation Engineering (RSETE), Nanjing, 13 June 2012, 14.

[15] Chen Jiong, Jia Haifeng, Yang Jian, et al. Primary exploration on monitoring of river pollution based on polarimetric coherence matrix［J］. Journal of remtoe sensing,(陈炯，贾海峰，杨健，等.基于极化相干矩阵的河流水质污染初探.遥感学报) 2011,15(5)： 10641078.

[16] Du Peijun, Xia Junshi, Zhang Wei, et al.Multiple Classifier System for Remote Sensing Image Classification： A Review［J］. Sensors, 2012,12(4)： 47644792.

[17] Yeung D, Kwok J, Fred A, et al. Structural, Syntactic, and Statistical Pattern Recognition［M］. Berlin Heidelberg, Springer 2006, 1939.

[18] Du Qian, Xia Junshi, Chanussot J, et al. Hyperspectral remote sensing image classification based on the integration of support vector machine and random forest［C］. Geoscience and Remote Sensing Symposium(IGARSS), Munich, 2227 July 2012, 174177.

[19] Du Peijun, Chen Yu, Xia Junshi, et al. A novel remote sensing image classification scheme based on data fusion, multiple features and ensemble learning［J］. J Indian Soc Remote sensing, 2013,41(2)： 213222

[20] Zhang Caiyu, Xie Zhixiao. Data fusion and classifier ensemble techniques for vegetation mapping in the coastal everglades. Geocarto International, 2012(35)： 16.

刘培, 杜培军, 谭琨. 一种基于集成学习和特征融合的遥感影像分类新方法[J]. 红外与毫米波学报, 2014, 33(3): 311. LIU Pei, DU Pei-Jun, TAN Kun. A novel remotely sensed image classification based on ensemble learning and feature integration[J]. Journal of Infrared and Millimeter Waves, 2014, 33(3): 311.

一种基于集成学习和特征融合的遥感影像分类新方法

关于本站 Cookie 的使用提示

全站搜索

一种基于集成学习和特征融合的遥感影像分类新方法

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索