**2.2 Conventional surface quality measurement and analysis techniques**

Vast amounts of work have been conducted in attempts to develop techniques to measure and evaluate surface texture in materials. These techniques generally fall into two distinct groups. The first is the hardware or method to measure surface texture data. The second is the analysis procedure to evaluate the surface texture. Numerous methods have been developed and researched for both the measurement and evaluation techniques. Measurement techniques normally fall into two distinct categories: contact and non-contact methods. It is beyond the scope of this chapter to discuss the surface measurement techniques that have been investigated in the past. The reader is referred to Lemaster (2004) for an overview of the various works on this topic. A general review of the optical techniques (and surface roughness techniques in general) is provided in several comprehensive reference works (Thomas, 1999; Whitehouse, 2011; Whitehouse, 1994; Thomas and King, 1977; and Riegel, 1993). The work conducted by the author on optical profilometry of wood and wood-based products can also be found in the literature (Lemaster, 2010, Lemaster, 2004; Lemaster 1997a, 1997b; Lemaster and Beall, 1996; Lemaster and DeVries, 1992; Lemaster and Dornfeld, 1983; Jouaneh, Lemaster, and Dornfeld, 1987; and DeVries and Lemaster, 1992).

The heart of any surface quality assessment system is the detector. The optical method used for the detector in this research is a variation of the reflectance method, whereby the positional change of the reflected laser light into the detector is correlated to changes in the test surface height. In this method, a laser spot is projected on the workpiece surface and the reflected light is focused on the surface of a lateral-effect photodiode. The change of the position of the reflected laser spot on the surface of the detector, a' is correlated to the vertical height change of the workpiece, a. By moving a workpiece beneath the detector and recording the change in the position of the laser spot, a two-dimensional surface profile is obtained that is very similar to that obtained by the traditional stylus system (Figure 2). The resulting surface profile can then be analyzed according to traditional U.S. (ASME B461- 2009) and international standards (ISO 4287/1). This method is non-contact and capable of detection at high speed, and since it measures position changes of the reflected light and not spot intensity, it is relatively insensitive to color changes of the workpiece.

Fig. 2. Schematic of optical profilometer

The Use of the Wavelet Transform to Extract

frequency domains were analyzed.

to wood machining is discussed below.

than the traditional phase correct Gaussian filter.

Additional Information on Surface Quality from Optical Profilometers 105

surface measurement. The use of standard surface descriptors based on time domain analysis is sufficient for some applications; however it does not provide information as to the periodicity of the surface characteristics or the nature or cause of the defects. Frequency is most often expressed as cycles per second known as Hertz (Hz). However, frequency can

As stated by Brock (1983), in the field of signal processing and analysis as applied to sound and vibration problems, the transformation of the signal from the time domain to the frequency domain is very common due to the ease with which the signal can be analyzed and characterized. Although this approach is not common in the field of surface quality analysis, the same benefits can be realized. The main advantage of frequency analysis is that it can reveal the dominant frequency components contained in the transducer signal. Ber and Braun (1968) showed that the frequency spectra resulting from the measurements on surfaces obtained by turning, grinding, and lapping are dissimilar. Raja and Radhakrishnan (1979) separated the roughness from the waviness component on a surface by using fast Fourier transform techniques. Staufert (1979) also used frequency domain analysis to separate periodic components from random components in the surface. In the literature an industry that has tended to use the power spectrum for surface quality analysis is that of road surface evaluation. In an article by Bruscella, Rouillard, and Sek (1999) a laser based optical profilometer was used to obtain a surface profile of the road. Both the time and

Work by Lemaster (1997b) has addressed the use of the frequency spectrum of the surface profile to detect "periodicity" within a surface profile. This approach is suitable because a surface profile is often composed of both random and periodic components. Under ideal cutter conditions, the tool produces evenly spaced cutter marks which occur periodically. In cases where the tool is not concentric, out of balance, or the workpiece is not properly held, the marks are unevenly spaced and vary in depth. More random defects often result from the detachment of material from the workpiece. The utility of simple frequency analysis is demonstrated, for several idealized (simulated) examples of surface quality issues relevant

Much work has been conducted on using wavelets in filtering or de-composing the surface profile. The category of interest here is the use of wavelets to separate these surface components. Much of the work discussing wavelets as applied to surface roughness are based on analyzing the gray scales of an image of the surface which is beyond the scope of this chapter and will not be discussed here but the reader is referred to Fricout et. al. (2002) for one discussion of this approach. Other works discussing wavelets and surface texture consists of multi-resolution decomposition of the surfaces including separating the error of form, waviness, roughness, and localized defects. Work by Khawaja (2011) demonstrated the insensitivity of the shape of the wavelet in its ability to decompose the components of a surface trace and obtain a standard roughness descriptor. While these works are very important in the complete understanding of surface texture analysis, it was not the main thrust of the topic in this chapter. In fact, the work by Lemaster (2004) found that this use of wavelets did indeed provide a means of removing the form of the surface texture that, in many cases, yielded superior filtering

also be expressed spatially such as cycles per unit length (cycles per inch).
