Convex restriction sets for CBERS-2 satellite image restoration

  • Authors:

  • J. P. Papa;N. D. A. Mascarenhas;L. M. G. Fonseca;K. Bensebaa

  • Affiliations:

  • UFSCAR- Federal University of São Carlos, São Carlos, Brazil;UFSCAR- Federal University of São Carlos, São Carlos, Brazil;INPE-National Institute for Space Research, São José dos Campos, Brazil;INPE-National Institute for Space Research, São José dos Campos, Brazil

  • Venue:
  • International Journal of Remote Sensing
  • Year:
  • 2008

Quantified Score

Hi-index 0.01

Visualization

Abstract

The main goal of this work is to develop a new efficient image restoration algorithm based on projections onto convex sets technique (POCS), which uses some a priori information about the images in the form of restriction sets. The proposed convex formulation used in this work is the prototype image constraints, which was obtained by the modified inverse filter, limited amplitude and by the modified row-action projection (MRAP) algorithm. A simulation experiment was performed using a high-resolution IKONOS image, which was blurred and decimated according to CBERS-2 CCD camera specifications. In order to allow quantitative analysis, the ISNR and the universal image quality index methodologies were applied. An original CBERS-2 CCD image was also used to evaluate the proposed restoration method. The restored images show a good visual performance, which can also be observed by the autocorrelation coefficients, which indicate high-frequency enhancement.