一种基于显微图像分析的光波导加工表面粗糙度估计方法

为进一步提高光波导器件加工的可靠性,降低废品率,提出一种非接触式基于共聚焦显微图像分析的光波导加工表面粗糙度估计方法.首先采用共聚焦相机对光波导感兴趣区域进行拍摄,提取非缺陷图像区域;其次,采用粗糙度测量设备测量器件的表面粗糙度,利用图像处理方法计算提取图像区域的纹理特征,并建立二者结果间对应关系;最后,一旦建立上述关系后,便可采用相机拍摄的图像去估计待测光波导器件表面的粗糙度指标.与传统采用精密光学仪器测量光波导表面粗糙度的方法相比,本文方法不需要在线使用光学测量仪器或昂贵的后处理商业软件,具有低成本、使用方便的特点.实际应用证明了本文所提方法的正确性和有效性.

A Surface Roughness Evaluation Method of Optic Waveguide Manufacture Based on Microscope Image Analysis

To improve the manufacture reliability of optic waveguide and decrease the unqualified product ratio to some extent, a noncontact surface roughness measurement method which uses the confocal microscopy is proposed. First, the confocal microscopy is utilized to capture the interested region of optic waveguide and choose the non-defect regions. Second, the texture features of these non-defect regions are computed, and their corresponding calculation results are compared with those results which are measured by the high precision surface roughness measurement device. Then the relationship between the image features and the measured surface roughness are built. Finally, once the relationship is built, the images captured by the confocal microscopy can be employed to estimate the surface roughness of optic waveguide. In contrast to the measurement method using high precision optic device, our low cost image based surface roughness measurement method can be utilized to implement an offline application rather than the online application of these devices. The practical product applications have proved the correction and effectiveness of this proposed method.