基于深度学习的激光熔覆层表面气孔识别研究

崔陆军; 刘亚轩; 郭士锐; 李海洋

光学技术, 2023, 49 (6): 673, 网络出版: 2023-12-05

基于深度学习的激光熔覆层表面气孔识别研究

Research on surface porosity recognition of laser cladding layer based on deep learning

论文大纲

崔陆军 ^1,2,*刘亚轩 ^1,2郭士锐 ^1,2李海洋 ^1,2

作者单位

¹ 中原工学院机电学院, 河南郑州 450007

² 郑州市激光增材制造技术重点实验室, 机械工业光学传感与测试技术重点实验室, 河南郑州 450000

激光熔覆语义分割熔覆层气孔深度学习串行注意力机制 laser cladding semantic segmentation stomata of cladding layer deep learning serial attention mechanism

摘要

为了解决熔覆层表面气孔识别技术中耗时且准确度不足的问题, 文章利用深度学习技术中的语义分割网络提出了基于U-net神经网络识别熔覆层表面气孔的2BNC-Unet神经网络。通过引入Batch Normalization层以及串联注意力机制(CBAM)合理部署在神经网络中, 选取交并比(IoU)与Dice系数作为网络的评价指标。研究结果表明: 在测试集中, 2BNC-Unet网络的交并比与Dice系数分别为86.96%、86.42%, 相比U-net神经网络分别提高了7.65%、4.73%。同时为了验证该网络的性能, 选用SegNet、2BNC-Unet与U-net神经网络进行对比实验, 结果表明2BNC-Unet的分割效果不仅优于SegNet和U-net网络, 而且熔覆层表面的气孔细节能够被完整地分割。在深度学习技术中2BNC-Unet的分割速度和准确度都有了显著地提高, 气孔的分割为熔覆层的性能分析提供了帮助。

Abstract

In order to solve the problems of time-consuming processes and insufficient accuracy in the surface porosity recognition technology of the cladding layer, A 2BNC-Unet neural network based on the U-Net neural network is proposed. The goal is to identify pores on the cladding layer's surface using semantic segmentation in deep learning technology. By introducing the Batch Normalization layer and the Convolutional Block Attention Module (CBAM) into the neural network in a reasonable manner, the Intersection over Union (IoU) and Dice coefficient were selected as evaluation indicators for the network. The results show that, in the test set, the intersection over union and Dice coefficient of the 2BNC-Unet network are 86.96% and 86.42%, respectively, which are 7.65% and 4.73% higher than those of the U-Net neural network. Additionally, to verify the performance of the network, comparative experiments were conducted using SegNet, 2BNC-Unet, and U-Net neural networks. The results demonstrate that the segmentation effect of 2BNC-Unet is not only better than that of SegNet and U-Net networks but also capable of completely segmenting the pore details on the cladding layer's surface. In the realm of deep learning technology, the segmentation speed and accuracy of 2BNC-Unet have been significantly improved, providing assistance in the performance analysis of cladding layers through pore segmentation.

参考文献

[1] 俞晓文, 姚喆赫, 张操棋, 等. 多维高频振动对激光熔覆成形几何形貌的影响研究［J］. 表面技术,2020,49(10):286-293+315.

[2] 侯锁霞, 赵江昆, 李强, 等. 对激光熔覆形成缺陷的影响因素的探究［J］. 材料导报,2022,36(S1):388-391.

[3] 刘钊鹏, 顾俊, 王健超, 等. WC-Co硬质合金涂层的激光熔覆工艺研究［J］. 应用激光,2021,41(4):715-723.

[4] Smurov I. Laser cladding and laser assisted direct manufacturing［J］. Surface and Coatings Technology,2008,202(18):4496-4502.

[5] 张浩, 练国富, 陈昌荣, 等. 激光熔覆曲面基体曲线轨迹成形内部质量与效率影响研究［J］. 激光与红外,2020,50(12):1433-1442.

[6] Cao C, Liu F, Tan H, et al. Deep learning and its applications in biomedicine［J］. Genomics, Proteomics & Bioinformatics,2018,16(1):17-32.

[7] 陈红花, 岑健, 刘溪, 等. 深度学习在化学流程工业故障诊断的研究进展［J］. 计算机工程与应用,2022,58(13):48-62.

[8] 于海洋, 景鹏, 张文涛, 等. 基于残差和注意力机制的道路裂缝检测U-Net改进模型［J/OL］. 计算机工程,(2022-9)/［2023-4］.https:∥kns.cnki.net/kcms/detail/31.1289.TP.20220930.1905.011.html

[9] 金澍, 关沫, 边玉婵, 等. 基于改进U-Net的遥感影像建筑物提取方法［J］. 激光与光电子学进展,2023,60(04):59-65.

[10] 何直蒙, 丁海勇, 安炳琪. 高分辨率遥感影像建筑物提取的空洞卷积E-Unet算法［J］. 测绘学报,2022,51(3):457-467.

[11] 赵兰枝, 史欣沅. 基于卷积神经网络的生物医学实体标准化研究［J］. 科技创新与应用,2022,12(15):30-35.

[12] 李琳辉, 钱波, 连静, 等. 基于卷积神经网络的交通场景语义分割方法研究［J］. 通信学报,2018,39(4):123-130.

[13] 刘丹, 刘学军, 王美珍. 一种多尺度CNN的图像语义分割算法［J］. 遥感信息,2017,32(1):57-64.

[14] He H, Wang S, Wang S, et al. A road extraction method for remote sensing image based on encoder-decoder network［J］. Journal of Geodesy and Geoinformation Science,2020,3(2):16-25

[15] 刘建伟, 刘俊文, 罗雄麟. 深度学习中注意力机制研究进展［J］. 工程科学学报,2021,43(11):1499-1511.

[16] 于营, 王春平, 付强, 等. 语义分割评价指标和评价方法综述［J］.计算机工程与应用,2023,59(06):57-69.

[17] 杨波, 梁伟. 深度神经网络模型构建及优化策略［J］. 计算机时代,2022,355(1):8-13+18.

[18] 饶竹一, 张云翔. 基于BiGRU和注意力机制的多标签文本分类模型［J］. 现代计算机,2020,673(01):31-35.

[19] 贵向泉, 张馨月, 李立. 高分辨率皮肤黑色素瘤图像的两阶段式分割算法［J/OL］. 计算机工程,(2023-05)/［2023-5］.

崔陆军, 刘亚轩, 郭士锐, 李海洋. 基于深度学习的激光熔覆层表面气孔识别研究[J]. 光学技术, 2023, 49(6): 673. CUI Lujun, LIU Yaxuan, GUO Shirui, Li Haiyang. Research on surface porosity recognition of laser cladding layer based on deep learning[J]. Optical Technique, 2023, 49(6): 673.

基于深度学习的激光熔覆层表面气孔识别研究

关于本站 Cookie 的使用提示

全站搜索

基于深度学习的激光熔覆层表面气孔识别研究

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索