针对光切显微镜测量材料表面粗糙度时,普通滤波算法不能将图像信号与噪声信号清楚分开这一问题,提出使用旋滤波方法对图像进行预处理,并与几种传统的滤波方法进行比较,同时讨论了滤波窗口的大小和切线方向采用的滤波方法对粗糙度测量结果的影响,得出的结论是:在光切法显微镜表面粗糙度图像的去噪滤波中,切线方向采用中值滤波的7×7窗口的旋滤波可以达到更好的效果。

Aimed the fact that the image signal and the noise can′t be separated distinctly by using the common filtering method when the material surface roughness is measured by light-section microscope. The spin filter method is proposed to preprocess the image. The roughness measurement results are compared with those measured with conventional methods. The influences of the size of filter window and the filtering method in the tangent direction on measurement results are discussed. The result is that using light-section microscope to remove noises of surface roughness image,the 7×7 spin filtering used on the tangent direction has a better effect.