Applications of integrated sensing and processing in spectroscopic imaging and sensing

Abstract

Integrated sensing and processing (ISP) encompasses the use of optical computing and adapted excitation signals to physically implement chemometric calculations in spectroscopic sensors for imaging. As data sets become larger and more complex with each emerging generation of hyperspectral imagers, the 'pixel-to-pupil' ratio increases at a rate faster than computing power can accommodate. In response to the need for faster and more efficient methods of processing, many analog solutions to the problem of high data dimensionality have emerged. The successful development of ISP has strong implications for military imaging, biosensing, spectroscopic imaging, and pharmaceutical process analytical technology (PAT). ISP developments in spectroscopy and PAT have emerged as alternatives to conventional Fourier transform infrared (FT-IR), near-infrared (NIR), IR, UV-visible, fluorescence, Raman, and acoustic-resonance spectrometry (ARS). Flourishing applications of ISP have demonstrated predictive ability equivalent to conventional approaches for sample differentiation and analyte quantification, in only a fraction of the time required for traditional spectrometric measurements.

Recent applications of Chemical Imaging to pharmaceutical process monitoring and quality control

“…PCA. This is the same as physically implementing chemometric calculations in spectroscopic systems [67]. Rapid analysis coupled with reduced data storage requirements indicate that the practical realisation of ICI systems as PATs is not far away.…”

Future Trends for PAT for Increased Process Understanding and Growing Applications in Biomanufacturing

“…Thus data acquisition can be minimized, while signal increased for a specifi c measurement, which can facilitate rapid measurements with high signalto -noise in systems. 34 Signal processing has applicability to many analytical methods, and serves as a complement to chemometric analysis, by focussing the measurement on the signals that are informative, rather than collecting data over an entire wavelength range. As data sets increase in size, ISPs will allow for their analysis without the need for taxing computational power.…”

Integrated Sensing and Processing Acoustic Resonance Spectrometry (ISP-ARS) for Sample Classification

“…In multivariate data where the concentrations of multiple species are changing simultaneously, calculation of standard figures of merit (FOM), such as dynamic range, limit of detection, selectivity, sensitivity, and signal-to-noise ratio is not a straightforward task [23,24]. The portion of the signal directly related to the analyte of interest is complicated by the signals from all other species and interferences present in the sample and sampling process.…”

“…The first way involves varying the excitation duration and/or intensity at particular frequencies such that each frequency is weighted according to its importance in the regression (encoding the excitation signal). The second way to accomplish the projection in analog hardware is by analog multiplication of the detected signal with an acoustic weight function after collection (post-collection) [24]. As illustrated in Fig.…”