基于多特征组合的光学元件表面疵病检测  

Surface defect detection of optical components based on multi-feature combination

作　　者：薛彬 吴志生[1] 孟庆森[2] XUE Bin;WU Zhisheng;MENG Qingsen(School of Materials Science and Engineering,Taiyuan University of Science and Technology,Taiyuan 030024,China;School of Mechanical and Electrical Engineering,Qingdao Binhai University,Qingdao Shandong 266555,China)

机构地区：[1]太原科技大学材料科学与工程学院,太原030024 [2]青岛滨海学院机电工程学院,山东青岛266555

出　　处：《激光杂志》2021年第12期108-113,共6页Laser Journal

基　　金：国家重点研发计划(No.2018YFA0707305);国家自然科学基金(No.51875384);山西省高校科技创新计划项目(No.STIP)(No.2019L0615);山西省青年科学基金(No.201901D211291);青岛滨海学院科技计划研究项目(No.2020KY04)。

摘　　要：采用目前方法对光学元件表面疵病检测时,由于没有利用光谱原理技术来获取光学元件的表面图像,导致检测方法的检测精度低、图像清晰度低、准确率低和判别正确率低,因此,提出基于多特征组合的光学元件表面疵病检测方法。首先利用光谱原理获取光学元件表面的图像,再结合高斯平滑曲线进行去噪预处理,采用Plessey算法进行角点的提取和匹配完成图像的拼接融合,之后再进行疵病的特征提取,计算疵病所占面积,从而完成光学元件表面疵病检测。实验结果表明,所提方法的检测精度高、图像清晰度高、准确率高和判别正确率高。When the current method is used to detect the surface defect of the optical element, because the spectrum principle technology is not used to obtain the surface image of the optical element, the detection method has low detection accuracy, low image clarity, low accuracy and low discrimination accuracy, so it is proposed Surface defect detection method of optical element based on multi-feature combination. First, use the spectrum principle to obtain the image of the surface of the optical element, then combine the Gaussian smooth curve for denoising preprocessing, use the Plessey algorithm to extract and match the corners to complete the image stitching and fusion, and then perform the feature extraction of the defect and calculate the defect Occupied area, thereby completing the detection of surface defects of optical components. Experimental results show that the proposed method has high detection accuracy, high image clarity, high accuracy and high discrimination accuracy.

关 键 词：光谱原理 高斯平缓曲线去噪 提取角点 图像拼接融合 疵病特征提取 

分 类 号：TN215[电子电信—物理电子学]