Polarimetric SAR image classification based on AdaBoost improved random forest and SVM

doi:10.7523/j.ucas.2021.0020

Abstract

Abstract: In order to improve the classification accuracy of polarimetric synthetic aperture radar (SAR) images, a two-level classification structure based on AdaBoost improved random forest (RF) and support vector machine (SVM) is proposed. Firstly, the AdaBoost improved RF (ADA_RF) is taken as the first-level classifier, which can assign weights according to the classification abilities of the decision trees. ADA_RF assigns high weights to strong decision trees. The first-level classifier can also assess the importance of input features and compute a ranking list. Feature selection can be conducted according to the list. The SVM classifier is trained with the selected features to predict the second-level classification result. Finally, the neighborhood voting method is used to fuse the results. The comparison experiments of AIRSAR polarization data shows that the classification structure can effectively improve the classification accuracy of polarimetric SAR images.

Key words: polarimetric SAR, AdaBoost, improved RF, two-level classifier, terrain classification

CLC Number:

TP753

ZHANG Zheng, LI Shiqiang. Polarimetric SAR image classification based on AdaBoost improved random forest and SVM[J]. Journal of University of Chinese Academy of Sciences, 2022, 39(6): 776-782.

References

[1] 张澄波. 综合孔径雷达: 原理、系统分析与应用(影印本)[M]. 北京: 科学出版社, 1989:1-3.
[2] Lee J S, Pottier E. 极化雷达成像基础与应用[M]. 洪文, 李洋, 尹嫱, 等译. 北京: 电子工业出版社，2013.
[3] Maitre H. 合成孔径雷达图像处理[M]. 孙洪，译. 北京: 电子工业出版社, 2005: 124.
[4] Chen S W, Tao C S. PolSAR image classification using polarimetric-feature-driven deep convolutional neural network[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(4): 627-631.DOI：10.1109/LGRS.2018.2799877.
[5] Liu H Y, Min Q, Sun C, et al. Terrain classification with polarimetric SAR based on deep sparse filtering network[C]//2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS). July 10-15, 2016, BeiJing, China. IEEE, 2016: 64-67.DOI:10.1109/IGARSS.2016.7729007.
[6] 张妙然, 刘畅. 基于特征筛选和二级分类的极化SAR建筑提取算法[J]. 中国科学院大学学报, 2018, 35(1): 89-95.DOI:10.7523/j.issn.2095-6134.2018.01.012.
[7] 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:10.1016/j.isprsjprs.2015.03.002.
[8] Salehi M, Sahebi M R, Maghsoudi Y. Improving the accuracy of urban land cover classification using radarsat-2 PolSAR data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(4): 1394-1401.DOI:10.1109/JSTARS.2013.2273074.
[9] 陈伟民, 张凌, 宋冬梅, 等. 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.
[10] Sun X, Song H J, Wang R, et al. High-resolution polarimetric SAR image decomposition of urban areas based on a POA correction method[J]. Remote Sensing Letters, 2018, 9(4): 363-372.
[11] Kumar A, Panigrahi R K. Entropy based reconstruction technique for analysis of hybrid-polarimetric SAR data[J]. IET Radar, Sonar & Navigation, 2019, 13(4): 620-626.DOI:10.1049/jet-rsn.2018.5338.
[12] Gola J, Webel J, Britz D, et al. Objective microstructure classification by support vector machine (SVM) using a combination of morphological parameters and textural features for low carbon steels[J]. Computational Materials Science, 2019, 160: 186-196.
[13] Zhao L J, Zhou X G, Kuang G Y. Building detection from urban SAR image using building characteristics and contextual information[J]. EURASIP Journal on Advances in Signal Processing, 2013, 2013(1): 56.
[14] Dekker R J. Texture analysis and classification of ERS SAR images for map updating of urban areas in The Netherlands[J]. IEEE Transactions on Geoscience and Remote Sensing, 2003, 41(9): 1950-1958.DOI:10.1109/TGRS.2013.814628.
[15] 陈刚, 王宏琦, 孙显. 基于核函数原型和自适应遗传算法的SVM模型选择方法[J]. 中国科学院研究生院学报, 2012, 29(1): 62-69.DOI:10.7523/j.issn.2095-6134.2012.1.009.
[16] Xiao D L, Liu C, W Q, et al. PolSAR image classification based on dilated convolution and pixel-refining parallel mapping network in the complex domain[EB/OL]. arXiv: 1909. 10783. (2020-01-20) [2021-03-08]. http://arxiv.org/abs/1909.10783.
[17] 滑文强, 王爽, 郭岩河, 等. 基于邻域最小生成树的半监督极化SAR图像分类方法[J]. 雷达学报, 2019, 8(4): 458-470.