Hyperspectral unmixing with <sub>q</sub> regularization

Research output: Contribution to journalArticlepeer-review

46 Citations (Scopus)

Abstract

Hyperspectral unmixing is an important technique for analyzing remote sensing images. In this paper, we consider and examine the lq, 0 ≤ q ≤ 1 penalty on the abundances for promoting sparse unmixing of hyperspectral data. We also apply a first-order roughness penalty to promote piecewise smooth endmembers. A novel iterative algorithm for simultaneously estimating the endmembers and the abundances is developed and tested both on simulated and two real hyperspectral data sets. We present an extensive simulation study where we vary both the SNR and the sparsity of the simulated data and identify the model parameters that minimize the reconstruction errors and the spectral angle distance. We show that choosing 0 ≤ q ≤ 1 can outperform the l1 penalty when the SNR is low or the sparsity of the underlying model is high. We also examine the effects of the imposing the abundance sum constraint using a real hyperspectral data set.

Original languageEnglish
Article number6776522
Pages (from-to)6793-6806
Number of pages14
JournalIEEE Transactions on Geoscience and Remote Sensing
Volume52
Issue number11
DOIs
Publication statusPublished - 2014

Other keywords

  • Blind signal separation
  • linear unmixing
  • roughness penalty
  • sparse regression

Fingerprint

Dive into the research topics of 'Hyperspectral unmixing with <sub>q</sub> regularization'. Together they form a unique fingerprint.

Cite this