小波滤波及奇异性分析在三维表面形貌评定中的应用

在机械加工领域，表面信息对于表面加工的指导作用至关重要。相对于表面二维信息，三维表面结构信息更能充分反映加工表面的实际特征，是热点研究问题。由于测量仪器的机械传动系统和测量环境等震动因素的存在，接触式测量方法得到的表面信号往往包含震动带入的误差信号，而这些信号存在其自身特点——奇异性特征。目前，小波变换是用于信号处理的有效工具。为透彻分析表面信号并剔除震动误差信号，本文介绍了通过小波变换对奇异信号进行检测的原理，并且给出了具体算法。在实验中利用该方法对表面测量信号进行预处理，并用小波滤波将表面结构分解得到粗糙度尺度轮廓，对其三维典型参数进行了计算，并与高斯滤波得到的粗糙度评定参数进行了比对。计算和比对表明小波奇异性分析是有效的，可得到更加合理的评定结果。

Application of wavelet filtering and singularity analysis to evaluation of surface roughness

In the domain of mechanical manufacture, the information from product surface was vitally important to supervising the surface manufacture. Compared with 2-Dimensional (2-D) information, the 3-Dimensional (3-D) information of surface texture as a hot research topic could more fully present the characteristics of machined surface. For the existence of vibration coming from mechanical transmission system and environment, the extracted surface by contact measurement sometimes included error ingredients, which presented characters of singularity. At present, wavelet transform was very useful to signal processing. In order to exclude the error ingredients mentioned above, a wavelet singularity analysis was adopted in the surface evaluation. Its principle and implementation algorithm were described in detail. In the experiment, the wavelet singularity analysis was used to preprocess the measured data first, and then the roughness profile was extracted by wavelet filter and Gaussian filter, the 3-D roughness results of which wre compared with each other. The computation and comparison proved that the algorithm was valid and more reasonable evaluation results could be obtained.