Thin infrared imaging systems through multichannel sampling

Abstract

The size of infrared camera systems can be reduced by collecting low-resolution images in parallel with multiple narrow-aperture lenses rather than collecting a single high-resolution image with one wide-aperture lens. We describe an infrared imaging system that uses a three-by-three lenslet array with an optical system length of 2.3 mm and achieves Rayleigh criteria resolution comparable with a conventional single-lens system with an optical system length of 26 mm. The high-resolution final image generated by this system is reconstructed from the low-resolution images gathered by each lenslet. This is accomplished using superresolution reconstruction algorithms based on linear and nonlinear interpolation algorithms. Two implementations of the ultrathin camera are demonstrated and their performances are compared with that of a conventional infrared camera.

Compact multi-aperture imaging with high angular resolution

“…For a system with α=1, employing a LWIR detector with 640×640 pixels at 25µm pitch, MA imaging with N=2 offers a length reduction without sacrificing angular resolution only for f'≤40mm. Recent MA implementations found in the literature correspond to α=1 and are subject to this constraint of low-resolution imaging [2][3][4][5][6].…”

“…(3) across channels the forward model of the system can be written as a classic restoration problem model, = + y Mx e (4) Due to its high dimensionality it is not feasible to uniquely invert the system matrix M to obtain an estimate of x, denoted by x , and instead iterative algorithms are preferred for inversion [22][23][24][25][26][27]. In this work the maximum likelihood estimator…”

“…When the angular resolution is limited by the size of the detector pixels, this can enable a significant reduction in imaging-system track length without sacrificing angular resolution. The use of a single detector array with an array of nominally identical lenses, as previously reported [2][3][4][5][6], limits the maximum aperture of the lens array to the width of the detector array and hence the benefit for MA imaging has previously been restricted to low-angular-resolution systems. We describe how the use of arrays of dissimilar lenses, some operating off-axis, but with an overall width wider than the detector array, enables the length-reduction benefit of MA imaging to be attained for high-resolution imaging.…”

“…In practice however, typical manufacturing tolerances tend to randomize sampling phase, maintaining the effectiveness of super-resolution with range [8]. On the second issue; the N-fold reduction in width of a lens aperture also reduces the diffraction-limited angular resolution by a factor N and thus MA imaging can maintain the angular resolution of a single-aperture imaging system only for small N. MA imaging have previously been used to improve compactness, but its application has been restricted to low-angular resolution imaging [2][3][4][5][6]. This restriction is associated with the use of arrays of identical lenslets, whereas we describe here how use of arrays of larger, but necessarily dissimilar lenslets, offers the potential of reduced track-length for high-angular-resolution imaging.…”

Super-resolution imaging using a camera array

“…If there are m lenslets across the width of the array, then a reduction in lens length by a factor of m is possible compared to a single lens of equal field-of-view, although the diffraction-limited resolution is limited to that of an aperture of width D∕m, where D is the width of the detector array. Recent advances employing multiaperture imaging include increased fieldof-view, implementation in the infrared, and improvements in signal processing, see [2][3][4][5][6] and references therein. None of these tackle the above angularresolution limit due to use of a single-detector array.…”