Using Wavelet Texture Analysis in Image-Based Classification and Statistical Process Control of Paper Surface Quality

Abstract

Paper formation (the level of homogeneity in the distribution of fibers on the surface of paper) is a key quality parameter for paper products, being currently monitored off-line, at low sampling rates relatively to the high production speeds achieved with modern paper machines. In this paper, we address the problem of conducting such monitoring activity on-line, in situ, using wavelet texture analysis (WTA) on raw images acquired with a especially design sensor. Our analysis shows that either a reduced set of features derived from WTA (suggested by an ANOVA analysis), or a low dimensional sub-space (with two dimensions, obtained using PCA or PLS-DA), enable an adequate separation of the several paper formation quality grades, meaning that we can indeed follow, on-line, the quality status of paper formation. Furthermore, we will also show how statistical process control (SPC) can be properly conducted using WTA features, in a simple and robust way.