高光谱影像子空间分析孤立森林异常目标探测方法

doi:10.11947/j.AGCS.2021.20200036

摘要/Abstract

摘要： 复杂的背景信息和高维冗余波段是影响高光谱遥感影像异常目标检测精度的重要因素。本文针对高光谱影像异常目标提取提出了一种子空间分析孤立森林探测方法。该方法不对背景做高斯分布假设，通过正交子空间分析增强输入特征影像中潜在异常目标与背景之间的对比度，通过主成分分析法降维来降低孤立森林算法带来的不确定性，运用了全局和局部结合的思想实现异常目标检测。在停机坪、海滩、港口和草地4个不同场景的高光谱影像上的试验结果表明，本方法的异常目标提取精度较经典方法取得了更好的结果。该方法不仅有效地处理了高光谱遥感影像的复杂背景和高维问题，还有效地利用了空间信息。

关键词: 高光谱影像, 异常目标检测, 孤立森林, 正交子空间, 主成分分析

Abstract: Since the anomalies are usually “rare and different” in the hyperspectral image scene, they tend to be more easily isolated from the background pixels by appropriate splitting criterion. In view of this, we propose a hyperspectral anomaly detection method based isolation forest (iForest) with subspace analysis. Firstly, orthogonal subspace background suppression and dimension reduction techniques were used to improve the reliability of the isolation tree-splitting criterion. Secondly, the iForest-based detection may produce a number of false alarms since the forest is constructed using the randomly selected pixels in the whole scene. In order to solve this problem, the initial anomaly detection map was refined by local iForest processing. Compared with original iForest method, our approach can not only handle high dimensional problem, but also make full use of the local information. The experiments demonstrate the AUC score have been significantly improved in our approach.

Key words: hyperspectral imagery, anomaly detection, isolation forest, orthogonal subspace, principal component analysis

中图分类号:

P237

黄远程, 薛园园, 李朋飞. 高光谱影像子空间分析孤立森林异常目标探测方法[J]. 测绘学报, 2021, 50(3): 416-425.

HUANG Yuancheng, XUE Yuanyuan, LI Pengfei. Subspace analysis isolation forest for hyperspectral anomaly detection[J]. Acta Geodaetica et Cartographica Sinica, 2021, 50(3): 416-425.

参考文献

[1] MATTEOLI S, DIANI M, CORSINI G. A tutorial overview of anomaly detection in hyperspectral images[J]. IEEE Aerospace and Electronic Systems Magazine, 2010, 25(7):5-28.
[2] ZHANG Yun. Smart photogrammetric and remote sensing image processing for very high resolution optical images-examples from the CRC-AGIP lab at UNB[J]. Journal of Geodesy and Geoinformation Science, 2019, 2(2):17-26. DOI:10.11947/j.JGGS.2019.0203.
[3] XIE Weiying, JIANG Tao, LI Yunsong, et al. Structure tensor and guided filtering-based algorithm for hyperspectral anomaly detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(7):4218-4230.
[4] DAI Yuchao, ZHANG Jing, HE Mingyi, et al. Salient object detection from multi-spectral remote sensing images with deep residual network[J]. Journal of Geodesy and Geoinformation Science, 2019, 2(2):101-110. DOI:10.11947/j.JGGS.2019.0211.
[5] HUYAN Ning, ZHANG Xiangrong, ZHOU Huiyu, et al. Hyperspectral anomaly detection via background and potential anomaly dictionaries construction[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(4):2263-2276.
[6] 李敏, 朱国康, 张学武, 等. 基于多孔径映射的高光谱异常检测算法[J]. 测绘学报, 2016, 45(10):1222-1230. DOI:10.11947/j.AGCS.2016.20160119. LI Min, ZHU Guokang, ZHANG Xuewu, et al. An anomaly detector based on multi-aperture mapping for hyperspectral data[J]. Acta Geodaetica et Cartographica Sinica, 2016, 45(10):1222-1230. DOI:10.11947/j.AGCS.2016.20160119.
[7] LU Xiaoqiang, WANG Yulong, YUAN Yuan. Graph-regularized low-rank representation for destriping of hyperspectral images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2013, 51(7):4009-4018.
[8] KRUSE F A, BOARDMAN J W, HUNTINGTON J F. Comparison of airborne hyperspectral data and EO-1 Hyperion for mineral mapping[J]. IEEE Transactions on Geoscience and Remote Sensing, 2003, 41(6):1388-1400.
[9] EISMANN M T, STOCKER A D, NASRABADI N M. Automated hyperspectral cueing for civilian search and rescue[J]. Proceedings of the IEEE, 2009, 97(6):1031-1055.
[10] IMANI M. RX anomaly detector with rectified background[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(8):1313-1317.
[11] KWON H, NASRABADI N M. Kernel RX-algorithm:a nonlinear anomaly detector for hyperspectral imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2005, 43(2):388-397.
[12] LING Qiang, GUO Yulan, LIN Zaiping, et al. A constrained sparse representation model for hyperspectral anomaly detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(4):2358-2371.
[13] SU Hongjun, WU Zhaoyue, DU Qian, et al. Hyperspectral anomaly detection using collaborative representation with outlier removal[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(12):5029-5038.
[14] YUAN Yuan, MA Dandan, WANG Qi. Hyperspectral anomaly detection via sparse dictionary learning method of capped norm[J]. IEEE Access, 2019, 7:16132-16144. DOI:10.1109/ACCESS.2019.2894590.
[15] LIU F T, TING Kaiming, ZHOU Zhihua. Isolation forest[C]//Proceedings of 2008 IEEE International Conference on Data Mining. Pisa:IEEE, 2009:413-422.
[16] LIU F T, TING Kaiming, ZHOU Zhihua. Isolation-based anomaly detection[J]. Acm Transactions on Knowledge Discovery from Data, 2012, 6(1):1-39.
[17] BANDARAGODA T R, TING Kaiming, ALBRECHT D, et al. Efficient anomaly detection by isolation using nearest neighbour ensemble[C]//2014 IEEE International Conference on Data Mining Workshop (ICDMW). Shenzhen:IEEE, 2014:698-705.
[18] ZHAO Rui, DU Bo, ZHANG Liangpei, et al. Beyond background feature extraction:an anomaly detection algorithm inspired by slowly varying signal analysis[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(3):1757-1774.
[19] 黄鸿, 石光耀, 段宇乐, 等. 加权空-谱联合保持嵌入的高光谱遥感影像降维方法[J]. 测绘学报, 2019, 48(8):1014-1024. DOI:10.11947/j.AGCS.2019.20180229. HUANG Hong, SHI Guangyao, DUAN Yule, et al. Dimensionality reduction method for hyperspectral images based on weighted spatial-spectral combined preserving embedding[J]. Acta Geodaetica et Cartographica Sinica, 2019, 48(8):1014-1024. DOI:10.11947/j.AGCS.2019.20180229.
[20] HARSANYI J C, CHANG C I. Hyperspectral image classification and dimensionality reduction:an orthogonal subspace projection approach[J]. IEEE Transactions on Geoscience and Remote Sensing, 1994, 32(4):779-785.
[21] 张兵. 高光谱图像处理与信息提取前沿[J]. 遥感学报, 2016, 20(5):1062-1090. ZHANG Bing. Advancement of hyperspectral image processing and information extraction[J]. Journal of Remote Sensing, 2016, 20(5):1062-1090.
[22] KHOUJ Y, DAWSON J, COAD J, et al. Hyperspectral imaging and k-means classification for histologic evaluation of ductal carcinoma in situ[J]. Frontiers in Oncology, 2018, 8:17. DOI:10.3389/fonc.2018.00017.
[23] BELHUMEUR P N, HESPANHA J P, KRIEGMAN D J. Eigenfaces vs. fisherfaces:recognition using class specific linear projection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1997, 19(7):711-720.
[24] OTSU N. A threshold selection method from gray-level histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1):62-66.
[25] KANG Xudong, ZHANG Xiangping, LI Shutao, et al. Hyperspectral anomaly detection with attribute and edge-preserving filters[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(10):5600-5611.
[26] 黄远程, 钟燕飞, 赵野鹤, 等. 联合盲分解与稀疏表达的高光谱图像异常目标检测[J]. 武汉大学学报(信息科学版), 2015, 40(9):1144-1150. HUANG Yuancheng, ZHONG Yanfei, ZHAO Yehe, et al. Joint blind unmixing and sparse representation for anomaly detection in hyperspectral image[J]. Geomatics and Information Science of Wuhan University, 2015, 40(9):1144-1150.
[27] WU Yuanfeng, LÓPEZ S, ZHANG Bing, et al. Approximate computing for onboard anomaly detection from hyperspectral images[J]. Journal of Real-Time Image Processing, 2019, 16(1):99-114.
[28] 马春笑, 黄远程, 胡荣明, 等. 多窗口融合判别子空间的高光谱图像异常检测[J]. 应用科学学报, 2019, 37(1):64-72. MA Chunxiao, HUANG Yuancheng, HU Rongming, et al. Discriminant subspace and multi-window fusion RX algorithm for hyperspectral image anomaly detection[J]. Journal of Applied Sciences, 2019, 37(1):64-72.
[29] ZHANG Yuxiang, KE Wu, DU Bo, et al. Independent encoding joint sparse representation and multitask learning for hyperspectral target detection[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(11):1933-1937.