非接触式表面粗糙度识别方法的研究

Study on the Method for Non-contact Roughness Detection

提出了一种机加工纹理非接触式表面粗糙度识别的方法。该方法用纹理周期度来描述机加工纹理的表面粗糙度,通过采集表面粗糙度标准样块的图像,用图像处理方法提取相应的纹理特征值(如纹理周期度)存入数据库,再用同样的方法获取未知粗糙度等级的零件表面图像的纹理周期特征值,并与库中粗糙度标准样块图像的纹理周期特征值进行比较,按照最近似的原则得到所求表面的粗糙度。该方法识别精度较高,具有对光照条件不同的免疫力。

In this paper, a novel non - contact approach to measure the roughness of machined surfaces is presented. The roughness of machined surfaces is described as texture - period in this method. Texture - period value of the image for the standard - roughness - piece is obtained with image processing, then it is saved in database. Texture -period value of the image for machined surfaces is obtained with the preceding method. Roughness of machined surfaces is captured after comparing texture - period value obtained ahead with one in database. This method is accurate relatively and it adapts different illumination.