IDENTIFICATION OF PLASTICS AMONG NONPLASTICS IN MIXED WASTE BY REMOTE-SENSING NEAR-INFRARED IMAGING SPECTROSCOPY .1. IMAGE IMPROVEMENT AND ANALYSIS BY SINGULAR-VALUE DECOMPOSITION

A near-IR camera has been installed in an experimental setup for real-time plastic identification. Singular value decomposition (SVD) has been used for qualitative analysis and substantial improvement of the measured multivariate images. The obtained score plots provided spatial correlations between different pixel structures caused by sample material on the one hand and image artifacts on the other. In this way, the score plots have been used as a tool to optimize the experimental setup and image quality. The improved images were offered to a new classification algorithm called multivariate image rank analysis, based on SVD, as described in part 2 of this series of articles, which follows in this issue (Wienke, D.; et al. Anal. Chem. 1995, 67, 3760).