摘要
传统光学零件表面缺陷检测方法以缺陷位置信息检测为主,在位置信息融合过程中存在信息遗漏问题,影响最终的检测精准度。因此,设计基于深度学习的光学零件表面缺陷检测方法。首先,提取光学零件表面缺陷特征,分析光学零件透镜中心成像情况,剔除中心误差导致的缺陷,保留光学零件表面缺陷特征。其次,基于深度学习检测光学零件表面缺陷细节尺度,获取零件缺陷的细节信息,并通过深度学习拟合缺陷特征。最后,进行实验分析。实验结果表明,该方法的检测精准度更高,优于对照组。
The traditional optical component surface defect detection method mainly focuses on defect location information detection,and there is a problem of information omission in the process of position information fusion,which affects the final detection accuracy.Therefore,a deep learning based surface defect detection method for optical components is designed.Firstly,extract the surface defect features of optical components,analyze the imaging situation of the lens center of optical components,eliminate defects caused by center errors,and retain the surface defect features of optical components.Secondly,based on deep learning,the detail scale of surface defects in optical components is detected,and the detailed information of part defects is obtained,and defect features are fitted through deep learning.Finally,conduct experimental analysis.The experimental results show that the detection accuracy of this method is higher and superior to the control group.
作者
刘伏涛
LIU Futao(China Aviation Industry Corporation Luoyang Electro Optic Equipment Research Institute,Luoyang Henan 471000,China)
出处
《信息与电脑》
2023年第21期64-66,共3页
Information & Computer
关键词
深度学习
光学零件
表面缺陷
检测方法
deep learning
optical parts
surface defects
detection method
