Reconstruction of light field with spectral information from compressive 2D projections

Abstract

A light field is an image with two spatial and two angular dimensions. Different acquisition systems have been used to capture light fields such as cameras and microlenses arrays. Compressive light field imaging is a recent technique that allows to recover a light field from just certain measurements sensed in a single 2-dimensional focal plane array (FPA). The essential optical element in this system is a binary coded aperture that blocks and unblocks the light rays before they impinge on the detector. Currently, it is studied the inclusion of a spectral dimension on the 4D light field with the aim to capture different spectral bands. This work presents a new acquisition model that includes the spectral dimension. The proposed acquisition model replaces the traditional binary coded apertures used in compressive light field by an array of optical filters which modulate the scene not only in the spatio — angular dimensions but spectrally as well. Simulations show that it is possible to reconstruct a light field with spectral information with only certain measures of the scene.