基于協峭度張量的高光譜圖像異常檢測

孟令博; 耿修瑞

doi:10.11999/JEIT180280

基于協峭度張量的高光譜圖像異常檢測

doi: 10.11999/JEIT180280 cstr: 32379.14.JEIT180280

孟令博,
耿修瑞^,

1.
中國科學院空間信息處理與應用系統(tǒng)技術重點實驗室 ??北京 ??100190
2.
中國科學院電子學研究所 ??北京 ??100190
3.
中國科學院大學 ??北京 ??100049

詳細信息

作者簡介:
孟令博：女，1989年生，博士生，研究方向為高光譜圖像特征提取及異常檢測

耿修瑞：男，1965年生，研究員，研究方向為高光譜圖像處理技術

通訊作者:
耿修瑞　xrgeng@mail.ie.ac.cn

中圖分類號: TP75
計量
- 文章訪問數: 1965
- HTML全文瀏覽量: 643
- PDF下載量: 71
- 被引次數: 0
出版歷程
- 收稿日期: 2018-03-26
- 修回日期: 2018-10-18
- 網絡出版日期: 2018-10-24
- 刊出日期: 2019-01-01

A Hyperspectral Imagery Anomaly Detection Algorithm Based on Cokurtosis Tensor

Lingbo MENG,
Xiurui GENG^,

1.
Key Laboratory of Technology in Geo-spatial Information Processing and Application System, Beijing 100190, China
2.
Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China
3.
University of Chinese Academy of Sciences, Beijing 100049, China

摘要

摘要:
高光譜圖像中的異常像元往往具有在圖像中出現的概率低和游離于背景數據云團之外的特點，如何“自動”確定這些異常像元是高光譜遙感圖像處理中的一個重要研究方向。經典的高光譜異常檢測方法一般從圖像的統(tǒng)計特性入手，廣泛應用的RXD異常檢測算法通過計算圖像的2階統(tǒng)計特征，可以直接給出異常點的分布情況，算法復雜度低，但缺點是沒有考慮到圖像的高階統(tǒng)計信息。基于獨立成分分析的異常檢測算法雖然考慮了高階統(tǒng)計量對異常點的敏感性，但需要反復迭代提取異常成分后，再對提取后的成分進行異常檢測。該文提出一種基于協峭度張量的異常檢測算法，該算法不需要事先提取異常成分，可以直接對觀測像元進行逐一檢測，從而給出異常點的分布情況。基于模擬數據和真實數據的實驗結果表明，該方法能夠在檢測出異常像元的同時更好地壓制背景信息、減小虛警率，從而提高異常檢測精度。
- 高光譜圖像 /
- 異常檢測 /
- 高階統(tǒng)計 /
- 協峭度張量
Abstract:
The abnormal pixels in hyperspectral images are often have the characteristics of low probability and scattered outside the background data cloud. How to automatically detect these abnormal pixels is an important research direction in hyperspectral imagery processing. Classical hyperspectral anomaly detection methods are usually based on statistical perspective. The RXD algorithm which is widely used can give the anomalies distribution directly through the second order statistical feature of the image, but the disadvantage is that it does not take into account the higher order statistics of the image. Anomaly detection algorithm based on Independent Component Analysis (ICA) considers the sensitivity of higher order statistics to outliers, but it needs iteration process to extract abnormal components first. And then the extracted components is used for anomaly detection. A method based on cokurtosis tensor for anomaly detection is proposed. This method does not need to extract anomaly components first. It can directly detect the observed pixels and give the distribution of abnormal pixels. Experiments results on both simulated and real data show that it can detect abnormal pixels while suppressing the background information better. Therefore, it can reduce false alarm rate and improve detection accuracy.
- Hyperspectral imagery /
- Anomaly detection /
- Higher-order statistical /
- Cokurtosis tensor

HTML全文

圖 1 數據在某個方向的偏度值(紅色線條長度)和峭度值(藍色線條長度)

下載: 全尺寸圖片幻燈片

圖 2 模擬數據各波段灰度圖

下載: 全尺寸圖片幻燈片

圖 3 異常檢測結果灰度圖

下載: 全尺寸圖片幻燈片

圖 4 COSD, RXD, COKD, KPCA-RXD算法的ROC曲線

下載: 全尺寸圖片幻燈片

圖 5 真實的高光譜圖像檢測結果

下載: 全尺寸圖片幻燈片

圖 6 4種異常檢測算法的ROC曲線

下載: 全尺寸圖片幻燈片

表 1 4種異常檢測算法的AUC

算法 AUC

COKD 0.9997
RXD 0.9934
COSD 0.9996
KPCA-RXD 0.9936

下載: 導出CSV

表 2 4種異常檢測算法的AUC

算法 AUC

COKD 0.9832
RXD 0.9779
COSD 0.9830
KPCA-RXD 0.9778

下載: 導出CSV

參考文獻(24)

Lü Qi, NIU Xin, DOU Yong, et al. Classification of hyperspectral remote sensing image using hierarchical local-receptive-field-based extreme learning machine[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(3): 434–438. doi: 10.1109/LGRS.2016.2517178

HEIDEN U, IWASAKI A, MULLER A, et al. Foreword to the special issue on hyperspectral remote sensing and imaging spectroscopy[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(9): 3904–3908. doi: 10.1109/JSTARS.2016.2610199

LI Wei and DU Qian. A survey on representation-based classification and detection in hyperspectral remote sensing imagery[J]. Pattern Recognition Letters, 2016, 83: 115–123. doi: 10.1016/j.patrec.2015.09.010

VERACINI T, MATTEOLI S, DIANI M, et al. Fully unsupervised learning of Gaussian mixtures for anomaly detection in hyperspectral imagery[C]. Ninth International Conference on Intelligent Systems Design and Applications, Pisa, Italy, 2009: 596–601. doi: 10.1109/ISDA2009220.

SALEM M B, ETTABAA K S, and BOUHLEL M S. Anomaly detection in hyperspectral images based spatial spectral classification[C]. International Conference on Sciences of Electronics, Technologies of Information and Telecommunications, Hammamet, Tunisia, 2017: 166–170. doi: 10.1109/SETIT.2016.7939860.

ZHANG Lili and ZHAO Chunhui. Hyperspectral anomaly detection based on spectral-spatial background joint sparse representation[J]. European Journal of Remote Sensing, 2017, 50(1): 362–376. doi: 10.1080/22797254.2017.1331697

THEILER J and ZIEMANN A K. Right spectrum in the wrong place: A framework for local hyperspectral anomaly detection[J]. Electronic Imaging, 2016, 2016(19): 1–9. doi: 10.2352/ISSN.2470-1173.2016.19.COIMG-160

CHIANG S S, CHANG C I, and GINSBERG I W. Unsupervised target detection in hyperspectral images using projection pursuit[J]. IEEE Transactions on Geoscience and Remote Sensing, 2001, 39(7): 1380–1391. doi: 10.1109/36.934071

WANG Lijing, GAO Kun, CHENG Xinman, et al. A hyperspectral imagery anomaly detection algorithm based on Gauss-Markov model[C]. Fourth International Conference on Computational and Information Sciences (ICCIS), Chongqing, China, 2012: 135–138. doi: 10.1109/ICCIS.2012.21.

耿修瑞. 高光譜遙感圖像目標探測與分類技術研究[D]. [博士論文], 中國科學院遙感與數字地球研究所, 2005: 81–91.

GENG Xiurui. Target detection and classification for hyperspectral imagery[D]. [Ph. D. dissertation], Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, 2015: 81–91.

TAGHIPOUR A and GHASSEMIAN H. Hyperspectral anomaly detection using attribute profiles[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(7): 1136–1140. doi: 10.1109/LGRS.2017.2700329

ZHANG Xing, WEN Gongjian, and DAI Wei. A tensor decomposition-based anomaly detection algorithm for hyperspectral image[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(10): 5801–5820. doi: 10.1109/TGRS.2016.2572400

TERREAUX E, OVARLEZ J P, and PASCAL F. Anomaly detection and estimation in hyperspectral imaging using random matrix theory tools[C]. IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, Cancun, Mexico, 2016: 169–172. doi: 10.1109/CAMSAP.2015.7383763.

XUN Lina and FANG Yonghua. Anomaly detection based on high-order statistics in hyperspectral imagery[C]. The Sixth World Congress on IEEE Intelligent Control and Automation, Dalian, China, 2006: 10416–10419. doi: 10.1109/WCICA.2006.1714044.

REN Hsuan and CHANG Yang Lang. A parallel approach for initialization of high-order statistics anomaly detection in hyperspectral imagery[C]. IEEE International Geosciences and Remote Sensing Symposium, Boston, USA, 2008: 1017–1020. doi: 10.1109/IGARSS.2008.4779170.

CARDOSO J F. Eigen-structure of the fourth-order cumulate tensor with application to the blind source separation problem[C]. International Conference on Acoustics, Speech, and Signal Processing, Albuquerque, USA, 1990: 2655–2658. doi: 10.1109/ICASSP.1990.116165.

OSAKO K, MORI Y, TAKAHASHI Y, et al. Fast convergence blind source separation based on frequency subband interpolation by null beamforming[C]. IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, New Paltz, USA, 2007. doi: 10.1109/ASPAA.2007.4392999.

GU Yanfeng, LIU Ying, and ZHANG Ye. A selective KPCA algorithm based on high-order statistics for anomaly detection in hyperspectral imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2008, 5(1): 43–47. doi: 10.1109/LGRS.2007.907304

成寶芝, 趙春暉, 王玉磊. 基于四階累積量的波段子集高光譜圖像異常檢測[J]. 光電子·激光, 2012, 23(8): 1582–1588.

CHENG Baozhi, ZHAO Chunhui, and WANG Yulei. Abnormal detection of hyperspectral images for band subsets based on fourth order cumulant[J]. Journal of Optoelectronics·Laser, 2012, 23(8): 1582–1588.

成寶芝. 基于光譜特性的高光譜圖像異常目標檢測算法研究[D]. [博士論文], 哈爾濱工程大學, 2014.

CHENG Baozhi. Abnormal target detection algorithm for hyperspectral images based on spectral characteristics[D]. [Ph.D. dissertation], Harbin Engineering University, 2014.

GENG Xiurui, SUN Kang, JI Luyan, et al. A high-order statistical tensor based algorithm for anomaly detection in hyperspectral imagery[J]. Scientific Reports, 2014, 4: 6869–6869. doi: 10.1038/srep06869

GENG Xiurui, JI Luyan, and SUN Kang. Principal skewness analysis: Algorithm and its application for multispectral/hyperspectral images indexing[J]. IEEE Geoscience and Remote Sensing Letters, 2014, 11(10): 1821–1825. doi: 10.1109/LGRS.2014.2311168

GENG Xiurui, JI Luyan, ZHAO Yongchao, et al. A small target detection method for the hyperspectral image based on Higher Order Singular Value Decomposition (HOSVD)[J]. IEEE Geoscience and Remote Sensing Letters, 2013, 10(6): 1305–1308. doi: 10.1109/LGRS.2013.2238504

孫康. 高光譜圖像波段選擇技術研究[D]. [博士論文], 中國科學院電子學研究所, 2015: 123–160.

SUN Kang. Research on band selection method for hyperspectral imagery[D]. [Ph.D. dissertation], Institute of Electrics, Chinese Academy of Sciences, 2015: 123–160.

施引文獻

資源附件(0)

訪問統(tǒng)計