Hyperspectral imagery super-resolution by spatial-spectral joint nonlocal similarity

Yongqiang Zhao, Jingxiang Yang, Jonathan Cheung Wai Chan

Research output: Contribution to journalArticlepeer-review

82 Scopus citations

Abstract

Hyperspectral (HS) super-resolution reconstruction is an ill-posed inversion problem, for which the solution from reconstruction constraint is not unique. To address this, an HS image super-resolution method is proposed to first utilize the joint regulation of spatial and spectral nonlocal similarities. We then fused the HS and panchromatic images with sparse regulation. With these two regulation terms, edge sharpness and spectrum consistency are preserved and noises are suppressed. The proposed method is tested with Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) and Hyperion images and evaluated by quantitative measures. The resulting enhanced images from the proposed method are superior to the results obtained by other well-known methods.

Original languageEnglish
Article number6684280
Pages (from-to)2671-2679
Number of pages9
JournalIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Volume7
Issue number6
DOIs
StatePublished - Jun 2014

Keywords

  • Hyperspectral (HS) imaging
  • nonlocal similarity
  • spectral unmixing
  • super-resolution

Fingerprint

Dive into the research topics of 'Hyperspectral imagery super-resolution by spatial-spectral joint nonlocal similarity'. Together they form a unique fingerprint.

Cite this