基于curvelet变换的光学器件表面特征识别

为了精确识别复杂精密光学器件在制造过程中由于工艺要求不可避免产生的缺陷特征,基于曲线波(Curvelet)变换优良的曲线特征识别和强大的稀疏表示能力,提出了一种基于曲线波变换的缺陷特征识别方法,该方法采用变换域特征分离,在空间域获得特征识别后的器件表面图像。仿真结果和实验结果证明,提出的方法比传统小波算法精度和准确度更好,运算速度也可以接受,因此可用于光学器件表面特征缺陷的在线检测。

Feature recognition of optical devices based on curvelet transform

In order to accurately identify the defect features which are inevitable in the manufacturing process of complex precision optical devices,based on the excellent curve feature recognition and strong sparse representation ability of curvelet transform,a defect feature recognition method based on curvelet transform is proposed. This method uses transform domain feature separation to obtain the identified component surface image in the spatial domain. The simulation results and experimental results show that the proposed method has better accuracy and accuracy than the traditional wavelet algorithm. The calculation speed is acceptable so that it can be used to detect surface feature defects of optical devices.