基于波段选择与学习字典的高光谱图像异常探测
|
摘要
针对高光谱影像数据中存在大量冗余,传统异常探测算法应用高光谱所有波段进行探测计算量巨大的问题,提出一种基于波段相似性线性预测与学习字典的异常探测算法。该算法首先通过对波段的相似性进行线性预测,找到最不相似的波段子集;然后,利用学习字典算法获得能够表征图像背景信息的背景字典,并通过低秩分解的算法将影像分解为低秩矩阵与稀疏矩阵;最后,使用经典RXD(Reed-X detector)探测算法对稀疏影像进行异常探测。实验结果表明,该算法可以在减少计算代价、保持波段原始信息不被破坏的同时,能够较好地实现了高光谱影像的异常探测。
With the large quantities of redundant information in the hyperspectral imagery,the traditional anomaly detection algorithm using the overall hyperspectral spectrum should consume a larger amount of computing time.Based on the linear prediction and learning dictionary,the authors put forward a novel algorithm.Compared with other low rank representation methods,the linear prediction method with the similarity of the band is utilized to find the least similar band subsets,and then the learning dictionary is implemented to obtain the learning dictionary which can represent the background information of the imagery.In addition,the imagery is divided into low rank matrix and sparse matrix via the low rank and decomposition.Finally,the traditional RXD(Reed-X detector)detection algorithm is utilized to detect the sparse image anomaly.Compared with other methods,the proposed method performs better with lower computational cost.Experimental results demonstrate that the selection of some bands including original information can achieve a good performance without corrupting the original information.It is a fine technique to apply to the hyperspectral imagery anomaly detection.
With the large quantities of redundant information in the hyperspectral imagery,the traditional anomaly detection algorithm using the overall hyperspectral spectrum should consume a larger amount of computing time.Based on the linear prediction and learning dictionary,the authors put forward a novel algorithm.Compared with other low rank representation methods,the linear prediction method with the similarity of the band is utilized to find the least similar band subsets,and then the learning dictionary is implemented to obtain the learning dictionary which can represent the background information of the imagery.In addition,the imagery is divided into low rank matrix and sparse matrix via the low rank and decomposition.Finally,the traditional RXD(Reed-X detector)detection algorithm is utilized to detect the sparse image anomaly.Compared with other methods,the proposed method performs better with lower computational cost.Experimental results demonstrate that the selection of some bands including original information can achieve a good performance without corrupting the original information.It is a fine technique to apply to the hyperspectral imagery anomaly detection.
引文
[1]钮宇斌,王斌.基于低秩表示和学习字典的高光谱图像异常探测[J].红外与毫米波学报,2016,35(6):731-740.Niu Y B,Wang B.Hyperspectral anomaly detection using lowrank representation and learned dictionary[J].Journal of Infrared and Millimeter Waves,2016,35(6):731-740.
[2]Reed I S,Yu X.Adaptive multiple-band CFAR detection of an optical pattern with unknown spectral distribution[J].IEEE Transactions on Acoustics Speech and Signal Processing,1990,38(10):1760-1770.
[3]Matteoli S,Diani M,Corsini G.A kurtosis-based test to efficiently detect targets placed in close proximity by means of local covariance-based hyperspectral anomaly detectors[C]//3rd Workshop on Hyperspectral Image and Signal Processing:Evolution in Remote Sensing(WHISPERS).Lisbon:IEEE,2011:1-4.
[4]Molero J M,Garzón E M,García I.Analysis and optimizations of global and local versions of the RX algorithm for anomaly detection in hyperspectral data[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2013,6(2):801-814.
[5]Taitano Y P,Geier B A,Bauer K W.A locally adaptable iterative RX detector[J].Eurasip Journal on Advances in Signal Processing,2010(1):1-10.
[6]Guo Q,Zhang B,Ran Q,et al.Weighted-RXD and linear filterbased RXD:Improving background statistics estimation for anomaly detection in hyperspectral imagery[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2014,7(6):2351-2366.
[7]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.
[8]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 B.Advancement of hyperspectral image processing and information extraction[J].Journal of Remote Sensing,2016,20(5):1062-1090.
[9]Xu Y,Wu Z,Li J.Anomaly detection in hyperspectral images based on low-rank and sparse representation[J].IEEE Transactions on Geoscience and Remote Sensing,2016,54(4):1990-2000.
[10]Li W,Du Q.Collaborative representation for hyperspectral anomaly detection[J].IEEE Transactions on Geoscience and Remote Sensing,2015,53(3):1463-1474.
[11]赵锐,杜博,张良培.一种基于核特征空间的鲁棒性高光谱异常探测方法[J].光子学报,2013,42(8):883-890.Zhao R,Du B,Zhang L P.An anomaly detection method for hyperspectral imagery in kernel feature space based on robust analysis[J].Acta Photonica Sinica,2013,42(8):883-890.
[12]张乐飞,张良培,陶大程.张量分类算法的遥感影像目标探测[J].遥感学报,2010,14(3):519-533.Zhang L F,Zhang L P,Tao D C.Tensor-based learning machine for remotely sensed image target detection[J].Journal of Remote Sensing,2010,14(3):519-533.
[13]彭波,张立福,张鹏,等.Cholesky分解的逐像元实时高光谱异常探测[J].遥感学报,2017,21(5):739-748.Peng B,Zhang L F,Zhang P,et al.A real-time sample-wise hyperspectral anomaly detection algorithm using Cholesky decomposition[J].Journal of Remote Sensing,2017,21(5):739-748.
[14]Huang S Y,Yeh Y R,Eguchi S.Robust kernel principal component analysis[J].Neural Computation.2009,21(11):3179-3213.
[15]Chen S Y,Yang S M,Kalpakis K,et al.Low-rank decompositionbased anomaly detection[C]//Proceedings of the SPIE Defense,Security,and Sensing.2013,8743.
[16]Liu G,Lin Z,Yan S,et al.Robust recovery of subspace structures by low-rank representation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(1):171-184.
[17]Niu Y,Wang B.A novel hyperspectral anomaly detector based on low-rank representation and learned dictionary[C]//IEEE International Geoscience and Remote Sensing Symposium(IGARSS).Beijing:IEEE,2016.
[18]Honeine P.Online kernel principal component analysis:A reducedorder model[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(9):1814-1826.
[19]Wang J,Chang C I.Independent component analysis-based dimensionality reduction with applications in hyperspectral image analysis[J].IEEE Transactions on Geoscience and Remote Sensing,2006,44(6):1586-1600.
[20]Weiland S,Van Belzen F.Singular value decompositions and low rank approximations of tensors[J].IEEE Transactions on Signal Processing,2010,58(3):1171-1182.
[21]Su M C,Chang H T.Fast self-organizing feature map algorithm[J].IEEE Transactions on Neural Networks,2000,11(3):721-733.
[22]哈斯巴干,马建文,李启青,等.多波段遥感数据的自组织神经网络降维分类研究[J].武汉大学学报(信息科学版),2004,29(5):461-465.Hasi B G,Ma J W,Li Q Q,et al.Dimension reduction of self-organized neural network classification for multi-band satellite data[J].Geomatics and Information Science of Wuhan University.2004,29(5):461-465.
[23]杜培军,王小美,谭琨,等.利用流形学习进行高光谱遥感影像的降维与特征提取[J].武汉大学学报(信息科学版),2011,36(2):148-152.Du P J,Wang X M,Tan K,et al.Dimensionality reduction and feature extraction from hyperspectral remote sensing imagery based on manifold learning[J].Geomatics and Information Science of Wuhan University.2011,36(2):148-152.
[24]李恒凯,吴立新,李发帅.面向土地利用分类的HJ-1 CCD影像最佳分形波段选择[J].遥感学报,2013,17(6):1572-1586.Li H K,Wu L X,Li F S,Optimal fractal band selection on HJ-1CCD image for land use classification[J].Journal of Remote Sensing,2013,17(6):1572-1586.
[25]Lu X Q,Li J L.A remote sensing images feature selection approach based on ant colony algorithm[C]//The 2nd International Conference on Industrial Mechatronics and Automation.Wuhan:IEEE,2010.
[26]Zhou S,Zhang J P,Su B K.Feature selection and classification based on ant colony algorithm for hyperspectral remote sensing images[C]//The 2nd International Congress on Image and Signal Processing.Tianjin:IEEE,2009.
[27]叶志伟,郑肇葆,万幼川,等.基于蚁群优化的特征选择新方法[J].武汉大学学报(信息科学版),2007,32(12):1127-1130.Ye Z W,Zheng Z B,Wan Y C,et al.A novel approach for feature selection based on ant colony optimization algorithm[J].Geomatics and Information Science of Wuhan University,2007,32(12):1127-1130.
[28]周爽.蚁群算法在高光谱图像降维和分类中的应用研究[D]哈尔滨:哈尔滨工业大学,2010.Zhou S.Research on the Application of Ant Colony Algorithm in the Dimentionality Reduction and Classification for Hyperspectral Image[D]Harbin:Harbin Institute of Technology,2010.
[29]Tan K,Li E,Du Q,et al.Hyperspectral image classification using band selection and morphological profile[C]//4th Workshop on Hyperspectral Image and Signal Processing:Evolution in Remote Sensing(WHISPERS).Shanghai:IEEE,2014.
[30]张兵,高连如.高光谱图像分类与目标探测[M].北京:科学出版社,2011.Zhang B,Gao L R.Hyperspectral Image Classification and Target Detection[M].Beijing:Science Press,2011.
[31]Wold S,Esbensen K,Geladi P.Principal component analysis[J].Chemometrics and Intelligent Laboratory Systems,1987,2(1-3):37-52.
[32]Li W,Du Q.Unsupervised nearest regularized subspace for anomaly detection in hyperspectral imagery[C]//IEEE International Geoscience and Remote Sensing Symposium(IGARSS).Melbourne:IEEE,2013.
[2]Reed I S,Yu X.Adaptive multiple-band CFAR detection of an optical pattern with unknown spectral distribution[J].IEEE Transactions on Acoustics Speech and Signal Processing,1990,38(10):1760-1770.
[3]Matteoli S,Diani M,Corsini G.A kurtosis-based test to efficiently detect targets placed in close proximity by means of local covariance-based hyperspectral anomaly detectors[C]//3rd Workshop on Hyperspectral Image and Signal Processing:Evolution in Remote Sensing(WHISPERS).Lisbon:IEEE,2011:1-4.
[4]Molero J M,Garzón E M,García I.Analysis and optimizations of global and local versions of the RX algorithm for anomaly detection in hyperspectral data[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2013,6(2):801-814.
[5]Taitano Y P,Geier B A,Bauer K W.A locally adaptable iterative RX detector[J].Eurasip Journal on Advances in Signal Processing,2010(1):1-10.
[6]Guo Q,Zhang B,Ran Q,et al.Weighted-RXD and linear filterbased RXD:Improving background statistics estimation for anomaly detection in hyperspectral imagery[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2014,7(6):2351-2366.
[7]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.
[8]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 B.Advancement of hyperspectral image processing and information extraction[J].Journal of Remote Sensing,2016,20(5):1062-1090.
[9]Xu Y,Wu Z,Li J.Anomaly detection in hyperspectral images based on low-rank and sparse representation[J].IEEE Transactions on Geoscience and Remote Sensing,2016,54(4):1990-2000.
[10]Li W,Du Q.Collaborative representation for hyperspectral anomaly detection[J].IEEE Transactions on Geoscience and Remote Sensing,2015,53(3):1463-1474.
[11]赵锐,杜博,张良培.一种基于核特征空间的鲁棒性高光谱异常探测方法[J].光子学报,2013,42(8):883-890.Zhao R,Du B,Zhang L P.An anomaly detection method for hyperspectral imagery in kernel feature space based on robust analysis[J].Acta Photonica Sinica,2013,42(8):883-890.
[12]张乐飞,张良培,陶大程.张量分类算法的遥感影像目标探测[J].遥感学报,2010,14(3):519-533.Zhang L F,Zhang L P,Tao D C.Tensor-based learning machine for remotely sensed image target detection[J].Journal of Remote Sensing,2010,14(3):519-533.
[13]彭波,张立福,张鹏,等.Cholesky分解的逐像元实时高光谱异常探测[J].遥感学报,2017,21(5):739-748.Peng B,Zhang L F,Zhang P,et al.A real-time sample-wise hyperspectral anomaly detection algorithm using Cholesky decomposition[J].Journal of Remote Sensing,2017,21(5):739-748.
[14]Huang S Y,Yeh Y R,Eguchi S.Robust kernel principal component analysis[J].Neural Computation.2009,21(11):3179-3213.
[15]Chen S Y,Yang S M,Kalpakis K,et al.Low-rank decompositionbased anomaly detection[C]//Proceedings of the SPIE Defense,Security,and Sensing.2013,8743.
[16]Liu G,Lin Z,Yan S,et al.Robust recovery of subspace structures by low-rank representation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(1):171-184.
[17]Niu Y,Wang B.A novel hyperspectral anomaly detector based on low-rank representation and learned dictionary[C]//IEEE International Geoscience and Remote Sensing Symposium(IGARSS).Beijing:IEEE,2016.
[18]Honeine P.Online kernel principal component analysis:A reducedorder model[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(9):1814-1826.
[19]Wang J,Chang C I.Independent component analysis-based dimensionality reduction with applications in hyperspectral image analysis[J].IEEE Transactions on Geoscience and Remote Sensing,2006,44(6):1586-1600.
[20]Weiland S,Van Belzen F.Singular value decompositions and low rank approximations of tensors[J].IEEE Transactions on Signal Processing,2010,58(3):1171-1182.
[21]Su M C,Chang H T.Fast self-organizing feature map algorithm[J].IEEE Transactions on Neural Networks,2000,11(3):721-733.
[22]哈斯巴干,马建文,李启青,等.多波段遥感数据的自组织神经网络降维分类研究[J].武汉大学学报(信息科学版),2004,29(5):461-465.Hasi B G,Ma J W,Li Q Q,et al.Dimension reduction of self-organized neural network classification for multi-band satellite data[J].Geomatics and Information Science of Wuhan University.2004,29(5):461-465.
[23]杜培军,王小美,谭琨,等.利用流形学习进行高光谱遥感影像的降维与特征提取[J].武汉大学学报(信息科学版),2011,36(2):148-152.Du P J,Wang X M,Tan K,et al.Dimensionality reduction and feature extraction from hyperspectral remote sensing imagery based on manifold learning[J].Geomatics and Information Science of Wuhan University.2011,36(2):148-152.
[24]李恒凯,吴立新,李发帅.面向土地利用分类的HJ-1 CCD影像最佳分形波段选择[J].遥感学报,2013,17(6):1572-1586.Li H K,Wu L X,Li F S,Optimal fractal band selection on HJ-1CCD image for land use classification[J].Journal of Remote Sensing,2013,17(6):1572-1586.
[25]Lu X Q,Li J L.A remote sensing images feature selection approach based on ant colony algorithm[C]//The 2nd International Conference on Industrial Mechatronics and Automation.Wuhan:IEEE,2010.
[26]Zhou S,Zhang J P,Su B K.Feature selection and classification based on ant colony algorithm for hyperspectral remote sensing images[C]//The 2nd International Congress on Image and Signal Processing.Tianjin:IEEE,2009.
[27]叶志伟,郑肇葆,万幼川,等.基于蚁群优化的特征选择新方法[J].武汉大学学报(信息科学版),2007,32(12):1127-1130.Ye Z W,Zheng Z B,Wan Y C,et al.A novel approach for feature selection based on ant colony optimization algorithm[J].Geomatics and Information Science of Wuhan University,2007,32(12):1127-1130.
[28]周爽.蚁群算法在高光谱图像降维和分类中的应用研究[D]哈尔滨:哈尔滨工业大学,2010.Zhou S.Research on the Application of Ant Colony Algorithm in the Dimentionality Reduction and Classification for Hyperspectral Image[D]Harbin:Harbin Institute of Technology,2010.
[29]Tan K,Li E,Du Q,et al.Hyperspectral image classification using band selection and morphological profile[C]//4th Workshop on Hyperspectral Image and Signal Processing:Evolution in Remote Sensing(WHISPERS).Shanghai:IEEE,2014.
[30]张兵,高连如.高光谱图像分类与目标探测[M].北京:科学出版社,2011.Zhang B,Gao L R.Hyperspectral Image Classification and Target Detection[M].Beijing:Science Press,2011.
[31]Wold S,Esbensen K,Geladi P.Principal component analysis[J].Chemometrics and Intelligent Laboratory Systems,1987,2(1-3):37-52.
[32]Li W,Du Q.Unsupervised nearest regularized subspace for anomaly detection in hyperspectral imagery[C]//IEEE International Geoscience and Remote Sensing Symposium(IGARSS).Melbourne:IEEE,2013.