Ultracompact Computational Spectroscopy and Hyperspectral Imaging

Spectroscopy is a salient practice for the identification and measurement of matter, but typical spectrometers are limited in both scope and efficiency with the current bulky benchtop apparatus. With space exploration becoming rampant, battlefield surveillance critical and forensic investigation real-time, the spectral analysis greatly warrants quick decision making in a compact, portable form factor without sacrificing performance.Targeting the infrared spectrum, where each substance has its own unique spectral fingerprint, this thesis proposes the design of such a compact, robust, low-cost IR spectrometer using our recently established method of Etalon Array Reconstructive Spectrometry (EARS), which facilitates even faster signal acquisition owing to the neat application of compressive sensing algorithms. In our method, we use an array of optical resonators (etalons) to uniquely encode the transmission spectrum of incident light, which later acts as a known sensing matrix for spectral reconstruction when recorded by a microbolometer array camera sensor. With a geometry that consists of no moving parts, inexpensive fabrication and added robustness of the versatile reconstruction algorithm, we endeavor to drive rapid and high-resolution spectroscopy. Furthermore, in an effort to overcome the inherent data deluge in hyperspectral imaging, we are encouraged to extend this study to devising handheld hyperspectral imagers with a high-speed, broadband imaging capability resolving thousands of spectral bands. Here, each pixel of the recorded image contains spectral information of the constituent object in the scene. We anticipate that our technology could easily and inexpensively integrate within the camera architectures in existing electronic systems