Home // SENSORDEVICES 2015, The Sixth International Conference on Sensor Device Technologies and Applications // View article


Macropixel Compressive Sensing Reconstruction of Spectral Images Sensed by Multispectral Filter Array-based Sensors

Authors:
Yuri Mejia
Fernando Rojas
Henry Arguello

Keywords: spectral images; MSFA-based sensors; compressive sensing; macropixel.

Abstract:
Spectral images are 3 dimensional data cubes of spectral information from a two-dimensional spatial scene. Traditional acquisition of these data cubes includes complex architectures that involves the use of prisms, tunable filters, and tunable illumination. Technical progress has allowed developing MSFA-based sensors in order to extrapolate the reconstruction of more bands than RGB cameras. However, reconstructing the spectral images with traditional mathematical methods following a least squares or demosaicing approach is unfeasible. Recently compressive sensing technique has been developed that allows reconstructing signals with few measurements than the traditional methods by using the sparse representation of the underlying signal. It is possible to exploit the capabilities of MSFA-based sensors selecting measurements subsets to form macropixels that have spectral information of a single reconstructed pixel. The macropixel size selection leads to a variable reconstructed spatial resolution preserving the filters spectral resolution. This paper presents a model for spectral images reconstruction from macropixels formed with MSFA-based sensor measurements using the principle of compressive sampling. This model selects subsets of the macropixels measurements following a downsampling matrix operation, therefore a reconstruction model is formulated by directly reconstruct a spectral image with the spectral resolution given by the number of filters. To verify the effectiveness of the reconstruction model measurements of the MSFA-based sensor for real spectral images are simulated. An ensemble of random dichroic filters is used. The macropixel compressive sensing reconstruction approach and the traditional scheme reconstruction are compared.

Pages: 158 to 163

Copyright: Copyright (c) IARIA, 2015

Publication date: August 23, 2015

Published in: conference

ISSN: 2308-3514

ISBN: 978-1-61208-426-8

Location: Venice, Italy

Dates: from August 23, 2015 to August 28, 2015