Development of an Optical Profilometer and the Related Advanced Signal Processing Methods for Monitoring Surface Quality of Wood Machining Applications

Abstract

The research described here provides the technology and theory to quantify surface quality for a variety of wood and wood-based products. This technology provides a means of monitoring trends in surface quality which can be used to discriminate between Agood@ products and Abad@ products (the methods described in this research are not intended to provide 'grading' of individual workpieces) as well as provide information to the machine operator as to the source of poor quality machined surfaces. The analysis can be done either on-line at industrial speeds or off-line as a periodic quality control tool. Although the surface quality can be quantifiably measured, the determination of the best feature from the surface profile (root mean square, peak amplitude, average wavelength, frequency content, Joint Time and Frequency Analysis (JTFA) and Wavelet Analysis results, etc.) for the quantification of surface Adefects@ is highly dependent on the application. This research consisted of three broad areas: (1) determination of an optimal hardware configuration for both laboratory and industrial surface scans of wood products, (2) determination of the optimal set of surface descriptors as well as the development of advanced signal processing techniques such as the wavelet transform to accurately describe the quality of a surface as well as provide information to the machine operator on the cause of the loss of surface quality, and (3) development of a software interface to distill the advanced signal processing techniques into a readily obtainable and readable format for the machine operator as well as provide assistance for process decisions.