QUALITY CONTROL OF POLYMER PRODUCTS THROUGH SPECTRAL IMAGING AND CHEMOMETRICS METHODS

摘要

Spectral imaging is increasingly used to develop rapid and non-invasive analytical sensors in a variety of fields. This paper discusses the ability of chemometrics methods, such as multivariate image analysis (MIA) and wavelet texture analysis (WTA), to extract meaningful information from spectral images, enabling the user to monitor subtle spatio-temporal variations in thin polymer materials. Three case studies are proposed to illustrate the method: (1) detecting subtle crystallinity variations across pure polymer films, (2) studying extrusion dynamics of wood/plastic composites making it possible to follow the mechanical properties on-line, and (3) predicting the mechanical properties of polymer blend films using both spatial and spectral features. These case studies show that spectral imaging can effectively be used identify local property variability and help in overall process control.