基于改进PSPNet的手机LCD屏幕表面缺陷检测  

Surface Defect Detection of Mobile Phone LCD Screen Based on Improved PSPNet

作　　者：肖彬 陈平华[1] XIAO Bin;CHEN Pinghua(School of Computer Science,Guangdong University of Technology,Guangzhou 510006,China)

机构地区：[1]广东工业大学计算机学院,广州510006

出　　处：《计算机测量与控制》2024年第9期36-43,共8页Computer Measurement &Control

基　　金：广东省重点领域研发计划项目(2023B1111050010,2020B0101100001)。

摘　　要：手机屏幕是智能手机的关键部件,其品质优劣直接影响到用户的使用体验;因此,手机屏幕缺陷检测成为工业生产中的重要环节;然而,手机LCD屏幕的表面缺陷检测目前还存在检测精确度低、模型参数较多等问题,无法满足实际工业生产需求;为了解决这些问题,对现有的缺陷检测算法和经典语义分割模型进行了研究,提出一种基于改进PSPNet的手机LCD屏幕表面缺陷检测模型;模型采用MobileNetV3作为特征提取网络,有效减少了模型参数;采用多尺度金字塔池化模块,进一步整合多尺度上下文信息,提高了模型的特征提取能力,有效应对屏幕图像中缺陷尺寸微小、边界模糊、相同缺陷尺寸差异较大的问题;同时,通过引入注意力机制,增强了模型的鲁棒性;实验结果表明,在SQ、Mura、TP、Line四种类型的手机LCD屏幕表面缺陷检测上,改进后的模型准确度明显优于基线模型。The screen of mobile phone is a key component of smartphones,the quality of screen directly affects the experience of users.Therefore,the screen defect detection of mobile phone has become an important part of industrial production.However,the surface defect detection of mobile liquid crystal display(LCD)screens has the problems of low detection accuracy and a large number of model parameters,which can not meet actual industrial production needs.To solve this problems,this paper studies existing defect detection algorithms and classical semantic segmentation models,and proposes an improved mobile phone LCD screen defect detection model based on pyramid scene parsing network(PSPNet).The model adopts the MobileNetV3 as the feature extraction network,which effectively reduces the model parameters.the multi-scale pyramid pooling module is used to further integrate the multi-scale contextual information,improving the feature extraction ability of the model.It also effectively addresses the issues of small defect sizes,blurred boundaries,and significant differences in same defect size in screen images.Meanwhile,the attention mechanism is introduced to increase the robustness of the model.Experimental results show that the accuracy of the improved model is significantly better than that of other traditional semantic segmentation models in the surface defect detection of four LCD screens:SQ,Mura,TP,and Line.

关 键 词：手机LCD屏幕 深度学习 缺陷检测 PSPNet 多尺度金字塔池化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]