Compressive polar spectral and polarization imaging
Date
2018
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Conventional digital imaging captures the desired image information directly
on an imaging sensor. When high dimensional imaging capability is required such as
spectral imaging and spectropolarimetric imaging, inecient sampling strategies are
usually involved, such as scanning and division of the focal plane array, resulting in
either poor real-time imaging capability or sensor resolution reduction. Compressive
sensing solves these problems via sampling of the high-dimensional image cubes below
the Nyquist sampling rate, and recovering the image signal with sparsity constraints.
In this dissertation, a newly proposed compressive spectral imaging and a compressive
spectropolarimetric imaging system are described. In particular, a novel polar-shaped
pixelated coding mask is proposed, optimized and implemented in spinning imaging
systems. Multi-snapshots are achieved through the coding mask rotation. By combin-
ing with a circular variable lter and a low-resolution monochrome sensor array, super-
resolution in both spatial and spectral dimensions is attained. Both the geometry and
the pattern of binary polar coded aperture are further optimized. A continuous imager
rotation model is developed and simulated to resolve the rotation blur. Finally, the
imaging system is tested with experimental measurements and reconstructions. On the
other hand, a compressive spectropolarimetric imager is developed based on a micro-
polarizer array, aiming at obtaining spectral linear Stokes images. Simulation shows
a blue-noise pattern of micro-polarizer distribution highly improves the polarization
reconstruction quality. A rotating double-amici prism further enables multi-snapshots.
Reconstructions of spectral Stokes images as well as the degree and angle of polarization
(DoP and AoP) images are then obtained from testbed experiments.
Description
Keywords
Applied sciences, Compressive sensing, Polar coded aperture, Compressive spectropolarimetric imager