Application Of Wavelet-Packet Analysis In Surface Roughness Measurement System

In the producing practices , people realize that the surface machining precision has remarkable influence on the capability and life-span of the accessory. With respect to the most of machines and instruments, especially the ones which run fast, need high assembly precision and good airproofing, the surface machining precision of their key accessories have to reach the relevant requirements and index, or the whole instrument would not run all right, even give birth to harsh accident.The surface machining precision of the accessories is denoted by relevant parameters, such as the parameters of surface roughness, the parameters of ripple, the parameters of shape error and so on. To obtain the precise surface appearance parameters of the accessories, people established a lot of Evaluating standard, and developed many instruments to measure the surface machining precision. With respect to whether contacts the measured surface or not, the measuring instruments are divided into two types: contacting and un-contacting. Recently,someone puts forward a optic-probe surface appearance measurement system. As an un-contacting system, it uses the feedback-information of the reaction of the focused laser beam and measured surface to measure the surface appearance parameters. This article mainly focuses on the signal-decomposition of this surface appearance measuring system .First, we introduced the meaning of surface appearance measurement, the definitions of relevant parameters, the common means of surface appearance measurement and their advantages and disadvantages. Meanwhile, we introduced the working theory and innovations of the surface appearance measurement system mentioned above. Then, with respect to the signal-decomposition problem of the measuring signal, we broke through traditional means, applied the theory of wavelet multi-definition analysis and wavelet-packet to this problem, advanced to decompose the measuring signal with the wavelet-packet analysis. This means averted the problem that the Fourier transform doesn't fit to the analysis of time-varying unsmooth signal , and the disadvantage that the wavelet analysis has low definition in the high frequencies. It can precisely separate the...