3D printing and additive manufacturing. 2024 Dec 16;11(6):e2045-e2060. doi: 10.1089/3dp.2023.0208 Q32.12025
Deep Learning-Based Automated Optical Inspection System for the Additive Manufacturing of Diamond Tools
基于深度学习的金刚石刀具增材制造自动光学检测系统 翻译改进
作者单位 +展开
作者单位
DOI: 10.1089/3dp.2023.0208 PMID: 39734731
摘要 中英对照阅读
使用增材制造技术有序排列金刚石磨粒的切割工具比传统金刚石工具更锋利且使用寿命更长。采用了一种可伸缩针状夹具,利用真空负压来吸收和放置有序排列的磨粒。然而,在长时间运行后,针孔磨损无法保证永久性地完全吸附磨粒。本文提出改进YOLOv5s模型来检测针孔上金刚石磨粒的吸附状态,以保持在增材制造过程中每个基体中金刚石磨粒的种植率。
首先,新增加的检测头提取更高层次的语义信息;其次,在模块中使用深度可分离卷积(DSC)+批量归一化+BRelu单元代替传统的卷积+批量归一化+Sigmoid激活函数来减少参数数量。将DSC引入Bottleneck1模块比直接应用瓶颈结构计算速度更快;最后,根据需要在适当位置添加坐标注意力机制以提高检测精度。
改进后的YOLOv5s模型实现了平均约19.6%的参数和每秒浮点运算量的减少。通过收集大量空穴及磨损空穴针孔的数据对该检查系统的性能进行了验证。与原始YOLOv5s相比,采用基于改进YOLOv5s模型的系统对一层金刚石磨粒进行检测的时间从6.35毫秒缩短到了5.06毫秒,并且检测精度高于98%。当吸附率低于95%时,会发出重新操作指令。该设备已连续运行一年,采用此增材制造设备生产的有序排列金刚石绿段的空穴率小于5%。
关键词:YOLOv5s;增材制造;深度学习;机器视觉检测系统;有序排列金刚石。
相关内容
-
Data of automated optical inspection of surface-mounted technology electronic production
表面贴装电子产品的自动光学检测数据
Korbinian Pfab et al.
Data in brief. 2024 Jan 30:53:110110.
-
Application of Additive Manufacturing and Deep Learning in Exercise State Discrimination
基于增材制造和深度学习的运动状态识别方法研究
Zhilong Zhao et al.
Sensors (Basel, Switzerland). 2025 Jan 10;25(2):389.
-
Deep Learning Applied to Defect Detection in Powder Spreading Process of Magnetic Material Additive Manufacturing
深度学习在磁性材料增材制造粉床铺粉缺陷检测中的应用研究
Hsin-Yu Chen et al.
Materials (Basel, Switzerland). 2022 Aug 17;15(16):5662.
-
Deep learning with mixup augmentation for improved pore detection during additive manufacturing
基于MixUp数据增强的深学习孔洞检测方法研究
Bulbul Ahmmed et al.
Scientific reports. 2024 Jun 11;14(1):13365.
-
Experimental investigation and development of a deep learning framework to predict process-induced surface roughness in additively manufactured aluminum alloys
实验研究和开发一种深度学习框架以预测铝合金增材制造工艺引起的表面粗糙度
Waqas Muhammad et al.
Welding in the world. 2023;67(4):897-921.
-
Automated Optical Inspection for Defect Identification and Classification in Actual Woven Fabric Production Lines
实际织物生产线上缺陷识别与分类的光学自动检测方法研究
Chung-Feng Jeffrey Kuo
Sensors (Basel, Switzerland). 2022 Sep 24;22(19):7246.
-
Assessment of Anisotropic Acoustic Properties in Additively Manufactured Materials: Experimental, Computational, and Deep Learning Approaches
基于实验、计算和深度学习的评价增材制造材料各向异性声学特性方法评估研究
Ivan Malashin et al.
Sensors (Basel, Switzerland). 2024 Jul 11;24(14):4488.
-
Broadband infrared confocal imaging for applications in additive manufacturing
宽带红外共焦显微成像及其在增材制造中的应用
Erin B Curry et al.
The Review of scientific instruments. 2022 Dec 1;93(12):123702.