基于改进Mobilenet-V3的纽扣电池注塑盖缺陷检测识别  

作　　者：吴泽强 张祺[1] 高学秋 

机构地区：[1]广东工业大学自动化学院,广东广州510006

出　　处：《工业控制计算机》2024年第11期42-44,共3页Industrial Control Computer

摘　　要：工业纽扣电池在生产过程中,不可避免会出现不良品。为解决人工挑拣效率低、漏检率高、成本高的问题,提出一种基于改进Mobilenet-V3网络的纽扣电池注塑盖缺陷检测模型。首先以Mobilenet-V3作为注塑盖缺陷特征提取的主干网络,对主干网络进行轻量化修剪,并引入了CBAM注意力机制替代Mobilenet-V3网络原有的SE注意力机制模块,增强模型的表征能力。In the production process of industrial button batteries,defective products will inevitably appear.In order to solve the problems of low efficiency,high omission rate and high cost of manual picking,a defect detection model of button battery cover based on improved Mobilenet-V3 network is proposed.Firstly,Mobilenet-V3 is used as the backbone network for defect feature extraction,and the backbone network is lightened and pruned,and CBAM attention mechanism is introduced to replace the SE attention mechanism module of Mobilenet-V3 network to enhance the representation ability of the model.

关 键 词：神经网络 Mobilenet-V3 CBAM注意力机制 图像识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]