Filtered gradient reconstruction algorithm for compressive spectral imaging

Yuri Mejia; Henry Arguello Sr.

doi:10.1117/1.OE.56.4.041306

9 November 2016 Filtered gradient reconstruction algorithm for compressive spectral imaging

Yuri Mejia, Henry Arguello Sr.

Author Affiliations +

Optical Engineering, Vol. 56, Issue 4, 041306 (November 2016). https://doi.org/10.1117/1.OE.56.4.041306

Abstract

Compressive sensing matrices are traditionally based on random Gaussian and Bernoulli entries. Nevertheless, they are subject to physical constraints, and their structure unusually follows a dense matrix distribution, such as the case of the matrix related to compressive spectral imaging (CSI). The CSI matrix represents the integration of coded and shifted versions of the spectral bands. A spectral image can be recovered from CSI measurements by using iterative algorithms for linear inverse problems that minimize an objective function including a quadratic error term combined with a sparsity regularization term. However, current algorithms are slow because they do not exploit the structure and sparse characteristics of the CSI matrices. A gradient-based CSI reconstruction algorithm, which introduces a filtering step in each iteration of a conventional CSI reconstruction algorithm that yields improved image quality, is proposed. Motivated by the structure of the CSI matrix, Φ, this algorithm modifies the iterative solution such that it is forced to converge to a filtered version of the residual Φ^Ty, where y is the compressive measurement vector. We show that the filtered-based algorithm converges to better quality performance results than the unfiltered version. Simulation results highlight the relative performance gain over the existing iterative algorithms.

Citation Download Citation

Yuri Mejia and Henry Arguello Sr. "Filtered gradient reconstruction algorithm for compressive spectral imaging," Optical Engineering 56(4), 041306 (9 November 2016). https://doi.org/10.1117/1.OE.56.4.041306

Published: 9 November 2016

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available