Demonstration of Visible and Near Infrared Raman Spectrometers and Improved Matched Filter Model for Analysis of Combined Raman Signals

摘要

Raman spectroscopy is a powerful analysis technique that has found applications in fields such as analytical chemistry, planetary sciences, and medical diagnostics. Recent studies have shown that analysis of Raman spectral profiles can be greatly assisted by use of computational models with achievements including high accuracy pure sample classification with imbalanced data sets and detection of ideal sample deviations for pharmaceutical quality control. The adoption of automated methods is a necessary step in streamlining the analysis process as Raman hardware becomes more advanced. Due to limits in the architectures of current machine learning based Raman classification models, transfer from pure to mixed sample analysis is not possible. This thesis presents the design, fabrication, and data collected from two different Raman spectrometers, a visible light system operating at 532 nm and a near infrared system operating at 785 nm. For each system, the optical design and operational theory of the main components will be explained. Data collected on each system will then be presented. Additionally, a learned matched filter computer model was developed to analyze Raman line profiles and can detect the signatures of multiple materials in a single data point. The presented model incorporates machine learning theory into the traditional matched filter model for higher probability of detection and much reduced probability of false alarm. The structure and operation of the model will be explained, and analysis of both real and simulated mixed-sample Raman spectra will be presented.