Semi-blind Hyperspectral Unmixing Using Nonnegative Matrix Factorization

Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 490)

Abstract

In hyperspectral imaging applications, spectral unmixing aims at identifying the constituent materials of a remotely sensed data and estimates its corresponding spectral signature for data exploitation. In this paper, the unmixing is primarily based on a linear mixture version in which every pixel is considered as a sum of definite number of absolutely clear spectra or endmembers, in accordance with means of abundance. Firstly, the number of endmembers in a given scene is determined using hyperspectral signal subspace identification by minimum error (Hysime) algorithm. Then, a vertex component analysis (VCA) method is used for unsupervised endmember extraction. Based on the observation that a negative reflectance is not possible, it is supportive and significant to constrain with nonnegativity. Thus, a nonnegative matrix factorization is applied for decomposing a given scene into its endmembers and abundance matrix. The successfulness of the researched technique is served using the simulated knowledge supported by USGS laboratory collected by the AVIRIS on mineral mining district, Nevada.

Keywords

Hyperspectral unmixing Nonnegative matrix factorization (NMF) Spectral signatures Blind source separation 

References

  1. 1.
    Keshava N, Mustard JF (2002) Spectral unmixing. IEEE Signal Process Mag 19(1):44–57CrossRefGoogle Scholar
  2. 2.
    Ma W et al (2014) Signal processing perspective on hyperspectral unmixing. IEEE Signal Process Mag 31(1):67–81CrossRefGoogle Scholar
  3. 3.
    Bayliss J, Gualtieri JA, Cromp R (1997) Analysing hyperspectral data with independent component analysis. Proc. SPIE 3240:133–143CrossRefGoogle Scholar
  4. 4.
    Comon P, Jutten C, Herault J (1991) Blind separation of sources, part II: problem statement. Signal Process 24:11–20CrossRefMATHGoogle Scholar
  5. 5.
    Boardman J (1993) Automating spectral unmixing of AVIRIS data using convex geometry concepts. In: Summaries 4th annual JPL Airborne geoscience workshop, vol 1, pp. 11–14. JPL Publication 93–26Google Scholar
  6. 6.
    Craig MD (1994) Minimum-volume transforms for remotely sensed data. IEEE Trans Geosci Remote Sens 32(1):99–109Google Scholar
  7. 7.
    Winter ME (1999) N-findr: an algorithm for fast autonomous spectral end-member determination in hyperspectral data. In Proceedings of the SPIE conference on imaging spectrometry V, pp 266–275Google Scholar
  8. 8.
    Aviris Cuprite Nevada Data set. [Online]. Available: http://aviris.jpl.nasa.gov/data/free_data.html

Copyright information

© Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  1. 1.Department of ECESSN College of EngineeringChennaiIndia

Personalised recommendations