Computational interferometry for hyperspectral imaging
(2019)
Files
Amirafshar_Moshtaghpour_thesis.pdf
Open Access - Adobe PDF
- 5.12 MB
Details
- Authors
- Supervisors
- Jacques, Laurent
- Abstract
- The idea of capturing hundreds to thousands images of an object, e.g., biological specimen, in different wavelengths has attracted decades of research under the name of HyperSpectral (HS) imaging. Since every chemical element in the universe has a unique signature in the spectral domain, obtaining such an HS data volume allows for characterizing the object elements, their concentration, and growth. This has made HS imaging an appealing tool in ubiquitous applications, e.g., chemistry, food science, agriculture, astronomy, and biology. Among a profusion of HS imaging techniques, the Fourier Transform Interferometry (FTI) has received a renewed interest for its high spectral resolution capability; a crucial criterion for, e.g., biomedical fluorescence spectroscopy. However, the effective resolution of the FTI is limited by the durability of biological elements when exposed to illuminating light, as over-exposed elements become unable to fluoresce. The main objective of this thesis is to investigate the possibilities to reduce the light exposure received by the observed object while preserving the spectral resolution of the reconstructed HS volumes. We propose novel computational interferometric techniques for acquisition of HS volumes based on the theory of compressive sensing with variable density sampling. In particular, we introduce three variations of the FTI system, i.e., Coded-Illumination FTI (CI-FTI), Structured-Illumination FTI (SI-FTI), and compressive Single-Pixel FTI (SP-FTI), based on the theory of compressive sensing. In these three contexts, our theoretical analysis addresses two main questions: (i) how to efficiently modulate light exposure temporally (in CI-FTI) or spatiotemporally (in SI-FTI and compressive SP-FTI) in order to minimize the light exposure imposed on the observed object? and (ii) how to recover high quality HS volumes from CI-FTI, SI-FTI, and compressive SP-FTI measurements?
- Affiliations
Citations
Moshtaghpour, A. (2019). Computational interferometry for hyperspectral imaging. https://hdl.handle.net/2078.5/92871