Development of on-line process analytical technologies based on spectral imaging and chemometrics for process and product quality control

该研究项目旨在开发基于光谱成像和化学计量学方法的新型在线过程分析传感器，用于测量化学过程和产品质量的可变性，这是传统传感器难以量化的。这些信息将有助于动态过程建模、监测、故障检测和诊断，并最终用于设计自动过程和质量控制方案。传感器将基于过程和产品的高光谱图像中包含的空间光谱信息而开发。最近从加拿大创新基金会（CFI）获得的两项资助中获得的UV-VIS-NIR光谱成像仪和高温相机将用于收集光谱图像。化学计量学方法，如主成分分析（PCA）和偏最小二乘回归（PLS）将用于有效地提取和分析过程和/或产品质量相关信息，这些信息包含在大量随机数据中，使每个光谱图像数据立方体。光谱传感器应该允许在线检测、表征和监测组成（有机和几种无机化合物）、形态和非均质材料和介质表面的外观（即空间分布）。

This research program aims at developing new on-line process analytical sensors based on spectral imaging and chemometrics methods for measuring variability in chemical processes and product quality that is difficult to quantify using traditional sensors. This information will be useful for dynamic process modelling, monitoring, fault detection and diagnosis, and ultimately for designing automatic process and quality control schemes. The sensors will be developed based on the spatial-spectral information contained in hyperspectral images of processes and products. The UV-VIS-NIR spectral imagers and a high temperature camera recently obtained from two Canada Foundation for Innovation (CFI) grants will be used for collecting spectral images. Chemometrics methods such as Principal Component Analysis (PCA) and Partial Least Squares Regression (PLS) will be used for efficient extraction and analysis of process and/or product quality relevant information contained in the massive amounts of stochastic data making each spectral image data cube. The spectral sensors should allow on-line detection, characterization, and monitoring of composition (organic and several inorganic compounds), morphology, and appearance across the surface of heterogeneous materials and media (i.e. spatial distributions).