Este artículo puede compartirse bajo la licencia CC BY-ND 4.0 y se referencia usando el siguiente formato: K. León, L. Galvis, H. Arguello, "Multiresolution-based reconstruction for compressive spectral video sensing using a spectral multiplexing sensor," Rev. UIS Ing., vol. 17, no. 1, pp. 209-216, 2018. Doi: https://doi.org/10.18273/revuin.v17n1-2018020 Vol. 17, no. 1, pp. 209-216, enero-junio 2018 Revista UIS Ingenierías ABSTRACTSpectral multiplexing sensors based on compressive sensing attempt to break the Nyquist barrier to acquire high spectral resolution scenes. Particularly, the colored coded aperture-based compressive spectral imager extended to video, or video C-CASSI, is a spectral multiplexing sensor that allows capturing spectral dynamic scenes by projecting each spectral frame onto a bidimensional detector using a 3D coded aperture. Afterwards, the compressed signal reconstruction is performed iteratively by finding a sparse solution to an undetermined linear system of equations. Even though the acquired signal can be recovered from much fewer observations by an ℓ 2 − ℓ 1 -norm recovery algorithm than using conventional sensors, the reconstruction exhibits diverse challenges originated by the temporal variable or motion. The motion during the reconstruction produces artifacts that damages the entire data. In this work, a multiresolution-based reconstruction method for compressive spectral video sensing is proposed. In this way, it obtains the temporal information from the measurements at a low computational cost. Thereby, the optimization problem to recover the signal is extended by adding temporal information in order to correct the errors originated by the scene motion. Computational experiments performed over four different spectral videos show an improvement up to 4dB in terms of peak-signal to noise ratio (PSNR) in the reconstruction quality using the multiresolution approach applied to the spectral video reconstruction with respect to the traditional inverse problem. KEYWORDS:Multiresolution reconstruction; compressive spectral video; optimization. RESUMENLos sensores de multiplexación espectral basados en muestreo compresivo intentan romper la barrera de Nyquist para adquirir escenas de alta resolución espectral. Particularmente, el sistema de imágenes espectrales de única captura basado en aperturas codificadas de color extendido a vídeo, o video -CCASSI, es un sensor de multiplexación espectral que permite la adquisición de imágenes espectrales dinámicas proyectando cada fotograma espectral sobre un detector bidimensional usando un apertura de codificación 3D. Posteriormente, la reconstrucción de la señal 210 K. León, L. Galvis, H. Arguello comprimida se realiza iterativamente encontrando una solución escasa a un sistema lineal de ecuaciones indeterminado. Si bien la señal adquirida puede ser recuperada desde un algoritmo basado en la norma l_2 − l_1, con muchas menos observaciones en comparación a los sistemas convencionales, dicha reconstrucción presenta diversos desafíos originados p...

ABSTRACT:In the last decade, spatio -angular (light field) acquisition systems have advanced due to the inclusion of coded apertures in the optical path. These coded apertures, modulate the light, encoding the information before being captured. Traditionally, these coded apertures are binary, i.e. block and unblock the light rays in the spatial dimension, capturing sparse information of the scene. In this work, the binary coded aperture is replaced by a colored coded aperture which modulates the source not only spatially but spectrally. Thereby, it is possible to capture light fields in multiple wavelengths yielding high spectral resolution. The spectral information provides many features of a scene in different wavelengths, these features are not present in the visible range of the electromagnetic spectrum. In this paper, an algorithm that simulates the light field sampling with colored coded apertures is proposed. The proposed algorithm, exploits the redundant information of the scene based on the compressive sensing theory thus, capturing just a sparse signal. The multidimensional image can be recovered from the underlying signal through a reconstruction algorithm. Several simulations show the quality of the multispectral light field reconstructions. The PSNR (Peak Signal to Noise Ratio) values obtained for the reconstructions are around 25 dB.