基于自适应回溯联合正交匹配追踪的高光谱图像解混
|
摘要
为提高高光谱图像的解混精度,针对联合匹配追踪(SMP)和联合正交匹配追踪(SOMP)算法在端元选择机制中存在的非最优问题,提出一种自适应回溯联合正交匹配追踪算法。对高光谱图像进行分块处理,通过初步测试选择每个分块中合适的端元加入端元支撑集,利用终极测试对支撑集中的端元进行检验,删除其中的错误端元,选取分块端元支撑集的并集作为整幅图像的端元支撑集,并以此为依据进行最小二乘法丰度估计。实验结果表明,与传统凸优化的稀疏解混算法及SMP,SOMP等贪婪算法相比,该算法具有更高的解混精度。
In order to increase the accuracy of hyperspectral unmixing,an Adaptive Backtracked Simultaneous Orthogonal Matching Pursuit(ABSOMP)algorithm is proposed to solve the problem that endmembers selection criterion of Simultaneous Matching Pursuit(SMP)and Simultaneous Orthogonal Matching Pursuit(SOMP)is not optimal in the sense of minimizing the residual of the new approximation.It uses block-processing strategy to divide the whole hyperspectral image into several blocks.Some potential endmembers are selected and added to the estimated endmember set in each block,then ABSOMP incorporates a backtracking process to detect the previous chosen endmembers' reliability and deletes the unreliable endmembers from the estimated endmember set in each iteration.The endmembers picked in each block are associated as the endmember sets of the whole hyperspectral data.Finally,abundances are estimated using the whole hyperspectral data with obtained endmember sets.Experimental results show that the unmixing accuracy of ABSOMP algorithm is better than traditional convex optimization sparse unmixing algorithm and other greedy algorithms such as SMP and SOMP algorithm.
In order to increase the accuracy of hyperspectral unmixing,an Adaptive Backtracked Simultaneous Orthogonal Matching Pursuit(ABSOMP)algorithm is proposed to solve the problem that endmembers selection criterion of Simultaneous Matching Pursuit(SMP)and Simultaneous Orthogonal Matching Pursuit(SOMP)is not optimal in the sense of minimizing the residual of the new approximation.It uses block-processing strategy to divide the whole hyperspectral image into several blocks.Some potential endmembers are selected and added to the estimated endmember set in each block,then ABSOMP incorporates a backtracking process to detect the previous chosen endmembers' reliability and deletes the unreliable endmembers from the estimated endmember set in each iteration.The endmembers picked in each block are associated as the endmember sets of the whole hyperspectral data.Finally,abundances are estimated using the whole hyperspectral data with obtained endmember sets.Experimental results show that the unmixing accuracy of ABSOMP algorithm is better than traditional convex optimization sparse unmixing algorithm and other greedy algorithms such as SMP and SOMP algorithm.
引文
[1]童庆禧,张兵,郑兰芬.高光谱遥感——原理、技术与应用[M].北京:高等教育出版社,2006.
[2]黄远程.高光谱影像混合像元分解的若干关键技术研究[D].武汉:武汉大学,2010.
[3]Marian-Daniel I,Bioucas-Dias J M,Plaza A.Sparse Unmixing of Hyperspectral Data[J].IEEE Transactions on Geoscience and Remote Sensing,2011,49(6):2014-2039.
[4]方红,杨海蓉.贪婪算法与压缩感知理论[J].自动化学报,2011,37(12):1414-1421.
[5]Shi Zhenwei,Tang Wei,Zhana D,et al.Subspace Matching Pursuit for Sparse Unmixing of Hyperspectral Data[J].IEEE Transactions on Geoscience and Remote Sensing,2014,52(6):3256-3274.
[6]Bioucas-Dias J M,Plaza A.Hyperspectral Unmixinggeometrical,Statistical,and Sparse Regression-based Approaches[C]//Proceedings of IEEE International Geoscience and Remote Sensing Symposium.Vancouver,Canada:IEEE Press,2011:1135-1138.
[7]Candes E J,Tao T.Near-optimal Signal Recovery from Random Projections:Universal Encoding Strategies?[J].IEEE Transactions on Information Theory,2006,52(12):5406-5425.
[8]Marian-Daniel I,Bioucas-Dias J M,Plaza A.Collaborative Sparse Regression for Hyperspectral Unmixing[J].IEEE Transactions on Geoscience and Remote Sensing,2014,52(1):341-354.
[9]Chen Jie,Huo Xiaoming.Theoretical Results on Sparse Representations of Multiple-measurement Vectors[J].IEEE Transactions on Signal Process,2006,54(12):4634-4643.
[10]Iordache M D.A Sparse Regression Approach to Hyper-spectral Unmixing[D].Lisbon,Portugal:Technical Uni-versity of Lisbon,2011.
[11]Tang Wei,Shi Zhenwei,Wu Ying.Regularized Simulta-neous Forward Backward Greedy Algorithm for Sparse Unmixing of Hyperspectral Data[J].IEEE Transactions on Geoscience and Remote Sensing,2014,52(9):5271-5188.
[12]曹离然,郑挺.多频带信号压缩采样的回溯SOMP重构方法[J].电子信息对抗技术,2012,27(4):41-46.
[13]Huang Honglin,Makur A.Backtracking-based Matching Pursuit Method for Sparse Signal Reconstruction[J].IEEE Signal Processing Letters,2011,18(7):391-394.
[2]黄远程.高光谱影像混合像元分解的若干关键技术研究[D].武汉:武汉大学,2010.
[3]Marian-Daniel I,Bioucas-Dias J M,Plaza A.Sparse Unmixing of Hyperspectral Data[J].IEEE Transactions on Geoscience and Remote Sensing,2011,49(6):2014-2039.
[4]方红,杨海蓉.贪婪算法与压缩感知理论[J].自动化学报,2011,37(12):1414-1421.
[5]Shi Zhenwei,Tang Wei,Zhana D,et al.Subspace Matching Pursuit for Sparse Unmixing of Hyperspectral Data[J].IEEE Transactions on Geoscience and Remote Sensing,2014,52(6):3256-3274.
[6]Bioucas-Dias J M,Plaza A.Hyperspectral Unmixinggeometrical,Statistical,and Sparse Regression-based Approaches[C]//Proceedings of IEEE International Geoscience and Remote Sensing Symposium.Vancouver,Canada:IEEE Press,2011:1135-1138.
[7]Candes E J,Tao T.Near-optimal Signal Recovery from Random Projections:Universal Encoding Strategies?[J].IEEE Transactions on Information Theory,2006,52(12):5406-5425.
[8]Marian-Daniel I,Bioucas-Dias J M,Plaza A.Collaborative Sparse Regression for Hyperspectral Unmixing[J].IEEE Transactions on Geoscience and Remote Sensing,2014,52(1):341-354.
[9]Chen Jie,Huo Xiaoming.Theoretical Results on Sparse Representations of Multiple-measurement Vectors[J].IEEE Transactions on Signal Process,2006,54(12):4634-4643.
[10]Iordache M D.A Sparse Regression Approach to Hyper-spectral Unmixing[D].Lisbon,Portugal:Technical Uni-versity of Lisbon,2011.
[11]Tang Wei,Shi Zhenwei,Wu Ying.Regularized Simulta-neous Forward Backward Greedy Algorithm for Sparse Unmixing of Hyperspectral Data[J].IEEE Transactions on Geoscience and Remote Sensing,2014,52(9):5271-5188.
[12]曹离然,郑挺.多频带信号压缩采样的回溯SOMP重构方法[J].电子信息对抗技术,2012,27(4):41-46.
[13]Huang Honglin,Makur A.Backtracking-based Matching Pursuit Method for Sparse Signal Reconstruction[J].IEEE Signal Processing Letters,2011,18(7):391-394.