Futuristic greedy approach to sparse unmixing of hyperspectral data

Naveed Akhtar, Faisal Shafait, Ajmal Mian

    Research output: Contribution to journalArticle

    20 Citations (Scopus)

    Abstract

    Spectra measured at a single pixel of a remotely sensed hyperspectral image is usually a mixture of multiple spectral signatures (endmembers) corresponding to different materials on the ground. Sparse unmixing assumes that a mixed pixel is a sparse linear combination of different spectra already available in a spectral library. It uses sparse approximation (SA) techniques to solve the hyperspectral unmixing problem. Among these techniques, greedy algorithms suite well to sparse unmixing. However, their accuracy is immensely compromised by the high correlation of the spectra of different materials. This paper proposes a novel greedy algorithm, called OMP-Star, that shows robustness against the high correlation of spectral signatures. We preprocess the signals with spectral derivatives before they are used by the algorithm. To approximate the mixed pixel spectra, the algorithm employs a futuristic greedy approach that, if necessary, considers its future iterations before identifying an endmember. We also extend OMP-Star to exploit the nonnegativity of spectral mixing. Experiments on simulated and real hyperspectral data show that the proposed algorithms outperform the state-of-the-art greedy algorithms. Moreover, the proposed approach achieves results comparable to convex relaxation-based SA techniques, while maintaining the advantages of greedy approaches.
    Original languageEnglish
    Pages (from-to)2157-2174
    JournalIEEE Transactions on Geoscience and Remote Sensing
    Volume53
    Issue number4
    DOIs
    Publication statusPublished - Apr 2015

    Fingerprint

    pixel
    Pixels
    Stars
    Derivatives
    experiment
    Experiments
    material
    library
    state of the art

    Cite this

    @article{2c1417f3a6ad475c8d344b1a29edb945,
    title = "Futuristic greedy approach to sparse unmixing of hyperspectral data",
    abstract = "Spectra measured at a single pixel of a remotely sensed hyperspectral image is usually a mixture of multiple spectral signatures (endmembers) corresponding to different materials on the ground. Sparse unmixing assumes that a mixed pixel is a sparse linear combination of different spectra already available in a spectral library. It uses sparse approximation (SA) techniques to solve the hyperspectral unmixing problem. Among these techniques, greedy algorithms suite well to sparse unmixing. However, their accuracy is immensely compromised by the high correlation of the spectra of different materials. This paper proposes a novel greedy algorithm, called OMP-Star, that shows robustness against the high correlation of spectral signatures. We preprocess the signals with spectral derivatives before they are used by the algorithm. To approximate the mixed pixel spectra, the algorithm employs a futuristic greedy approach that, if necessary, considers its future iterations before identifying an endmember. We also extend OMP-Star to exploit the nonnegativity of spectral mixing. Experiments on simulated and real hyperspectral data show that the proposed algorithms outperform the state-of-the-art greedy algorithms. Moreover, the proposed approach achieves results comparable to convex relaxation-based SA techniques, while maintaining the advantages of greedy approaches.",
    author = "Naveed Akhtar and Faisal Shafait and Ajmal Mian",
    year = "2015",
    month = "4",
    doi = "10.1109/TGRS.2014.2356556",
    language = "English",
    volume = "53",
    pages = "2157--2174",
    journal = "IEEE Transactions on Geoscience and Remote Sensing",
    issn = "0196-2892",
    publisher = "IEEE, Institute of Electrical and Electronics Engineers",
    number = "4",

    }

    Futuristic greedy approach to sparse unmixing of hyperspectral data. / Akhtar, Naveed; Shafait, Faisal; Mian, Ajmal.

    In: IEEE Transactions on Geoscience and Remote Sensing, Vol. 53, No. 4, 04.2015, p. 2157-2174.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - Futuristic greedy approach to sparse unmixing of hyperspectral data

    AU - Akhtar, Naveed

    AU - Shafait, Faisal

    AU - Mian, Ajmal

    PY - 2015/4

    Y1 - 2015/4

    N2 - Spectra measured at a single pixel of a remotely sensed hyperspectral image is usually a mixture of multiple spectral signatures (endmembers) corresponding to different materials on the ground. Sparse unmixing assumes that a mixed pixel is a sparse linear combination of different spectra already available in a spectral library. It uses sparse approximation (SA) techniques to solve the hyperspectral unmixing problem. Among these techniques, greedy algorithms suite well to sparse unmixing. However, their accuracy is immensely compromised by the high correlation of the spectra of different materials. This paper proposes a novel greedy algorithm, called OMP-Star, that shows robustness against the high correlation of spectral signatures. We preprocess the signals with spectral derivatives before they are used by the algorithm. To approximate the mixed pixel spectra, the algorithm employs a futuristic greedy approach that, if necessary, considers its future iterations before identifying an endmember. We also extend OMP-Star to exploit the nonnegativity of spectral mixing. Experiments on simulated and real hyperspectral data show that the proposed algorithms outperform the state-of-the-art greedy algorithms. Moreover, the proposed approach achieves results comparable to convex relaxation-based SA techniques, while maintaining the advantages of greedy approaches.

    AB - Spectra measured at a single pixel of a remotely sensed hyperspectral image is usually a mixture of multiple spectral signatures (endmembers) corresponding to different materials on the ground. Sparse unmixing assumes that a mixed pixel is a sparse linear combination of different spectra already available in a spectral library. It uses sparse approximation (SA) techniques to solve the hyperspectral unmixing problem. Among these techniques, greedy algorithms suite well to sparse unmixing. However, their accuracy is immensely compromised by the high correlation of the spectra of different materials. This paper proposes a novel greedy algorithm, called OMP-Star, that shows robustness against the high correlation of spectral signatures. We preprocess the signals with spectral derivatives before they are used by the algorithm. To approximate the mixed pixel spectra, the algorithm employs a futuristic greedy approach that, if necessary, considers its future iterations before identifying an endmember. We also extend OMP-Star to exploit the nonnegativity of spectral mixing. Experiments on simulated and real hyperspectral data show that the proposed algorithms outperform the state-of-the-art greedy algorithms. Moreover, the proposed approach achieves results comparable to convex relaxation-based SA techniques, while maintaining the advantages of greedy approaches.

    U2 - 10.1109/TGRS.2014.2356556

    DO - 10.1109/TGRS.2014.2356556

    M3 - Article

    VL - 53

    SP - 2157

    EP - 2174

    JO - IEEE Transactions on Geoscience and Remote Sensing

    JF - IEEE Transactions on Geoscience and Remote Sensing

    SN - 0196-2892

    IS - 4

    ER -