基于刀具刃口显微图像的附着物去除网络

doi:10.19678/j.issn.1000-3428.0060856

计算机工程 ›› 2022, Vol. 48 ›› Issue (4): 247-254,261. doi: 10.19678/j.issn.1000-3428.0060856

基于刀具刃口显微图像的附着物去除网络

梁智滨¹, 赵文义², 李灵巧³, 杨辉华^1,3

1. 桂林电子科技大学电子工程与自动化学院, 广西桂林 541004;
2. 北京邮电大学人工智能学院, 北京 100876;
3. 桂林电子科技大学计算机与信息安全学院, 广西桂林 541004

收稿日期:2021-02-09 修回日期:2021-05-07 发布日期:2021-05-20
作者简介:梁智滨(1996—),男,硕士研究生,主研方向为测量技术与智能系统;赵文义,博士研究生;李灵巧,讲师、博士;杨辉华,教授、博士、博士生导师。
基金资助:
国家自然科学基金青年科学基金项目（61906050）；广西科技计划项目（AD19245202）。

Attachment Removal Network Based on Micro Image of Tool Edge

LIANG Zhibin¹, ZHAO Wenyi², LI Lingqiao³, YANG Huihua^1,3

1. College of Electronic Engineering and Automation, Guilin University of Electronic Technology, Guilin, Guangxi 541004, China;
2. College of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing 100876, China;
3. College of Computer and Information Security, Guilin University of Electronic Technology, Guilin, Guangxi 541004, China

Received:2021-02-09 Revised:2021-05-07 Published:2021-05-20

摘要/Abstract

摘要： 准确检测并去除刀具边缘粘连的附着物是刀具显微图像豁口检测领域的一个难题，目前仍存在刀具边缘恢复不完整、附着物去除失败等问题。提出一种附着物去除的ARNet网络，采用二值掩膜引导模块区分目标与背景特征，利用去除过程的学习模块提取递归过程中的时序信息，并通过自注意力精准分离模块中的编解码结构和自注意力机制，建立附着物特征在全局特征中的依赖关系，以精准去除附着物，从而整合特征并输出无附着物图像。从实际采集的刀具刃口显微图像中裁剪含附着物区域的图像并构成数据集，在此数据集上的实验结果表明，与PReNet网络相比，该网络的峰值信噪比提高了1.016 dB，交并比IOU提升了3.48%，参数量和计算量分别减少了86.5%、90.9%，能够精确聚焦附着物区域，完整地还原刀具的真实边缘，提高了豁口检测准确率，且增强了刀具豁口高精度检测系统的稳定性和可靠性。

关键词: 机器视觉, 自注意力机制, 刀具刃口, 显微图像, 附着物检测

Abstract: Detecting and removing the adhesion on a tool edge accurately is a difficult problem in the field of tool micro-image notch detection.Some problems remain, such as incomplete tool edge restoration and failure of attachment removal.To solve the problems, an Attachment-Removal Network(ARNet) is proposed.The Binary Mask Guidance Module(BMGM) is used to distinguish the target and background features, and the learning module of the removal process is used to extract the timing information in the recursive process.Through the encoder-decoder structure and Self-Attention(SA) mechanism in the Self-Attention Refined Separation Module(SARSM), the dependency of attachment features in the global features is established to remove attachments accurately;thus, the features are integrated, and the attachment-free image is output.The image with the attachment area is cut from the actual collected micro image of the tool edge to form a data set.The results show that, compared with PReNet, the Peak Signal-to-Noise Ratio(PSNR) and Intersection Over Union(IOU) of the proposed method are improved by 1.016 dB and 3.48%, respectively.In addition, 86.5% of the parameters and 90.9% of the calculations are reduced.This method focuses on the attachment area accurately, restores the real edge of the tool completely, improves the accuracy of gap detection, and enhances the stability and reliability of the high-precision detection system for the tool gap.

Key words: machine vision, self-attention mechanism, knife tool edge, microscopic image, attachment detection

中图分类号:

TP391.4

梁智滨, 赵文义, 李灵巧, 杨辉华. 基于刀具刃口显微图像的附着物去除网络[J]. 计算机工程, 2022, 48(4): 247-254,261.

LIANG Zhibin, ZHAO Wenyi, LI Lingqiao, YANG Huihua. Attachment Removal Network Based on Micro Image of Tool Edge[J]. Computer Engineering, 2022, 48(4): 247-254,261.

http://www.ecice06.com/CN/Y2022/V48/I4/247

图/表 11

20230131202216

20230131202219

20230131202223

20230131202227

20230131202230

20230131202233

20230131202236

20230131202239

20230131202242

20230131202245

20230131202248

参考文献

[1] 颜振翔.刀具豁口高精度自动检测装置关键技术研究及实现[D].桂林:桂林电子科技大学, 2019. YAN Z X.Research and implementation of key technoloay for high precision automatic detecting device for tool flaw[D].Guilin, Guangxi:Guilin University of Electronic Technology, 2019.(in Chinese)
[2] 颜振翔, 王寒迎, 石齐双, 等.基于区域蛙跳搜索与轮廓匹配的显微图像拼接[J].激光与光电子学进展, 2019, 56(15):65-72. YAN Z X, WANG H Y, SHI Q S, et al.Microscopic image stitching based on regional frog leaping search algorithm and image contour matching[J].Laser and Optoelectronics Progress, 2019, 56(15):65-72.(in Chinese)
[3] 栗琳, 王仲, 蔡振兴, 等.基于目标轮廓的附着物定位与剔除方法[J].光电工程, 2012, 39(5):45-51. LI L, WANG Z, CAI Z X, et al.A method of location and elimination of foreign matters based on navigation contour[J].Opto-Electronic Engineering, 2012, 39(5):45-51.(in Chinese)
[4] GUO X J, YANG X Y, YU Z J.Foreign object debris detection on the runway based on wavelet method[J].Applied Mechanics and Materials, 2013, 427(3):1658-1661.
[5] 张辉, 王耀南, 周博文.基于机器视觉的液体药品异物检测系统研究[J].仪器仪表学报, 2009, 30(3):548-553. ZHANG H, WANG Y N, ZHOU B W.Research on foreign substance detection system for medicinal solution based on machine vision[J].Chinese Journal of Scientific Instrument, 2009, 30(3):548-553.(in Chinese)
[6] MI C, CHEN K, ZHANG Z W.Research on tobacco foreign body detection device based on machine vision[J].Transactions of the Institute of Measurement and Control, 2020, 42(2):2857-2871.
[7] ISKANDAR D N F A, LING N J, FAUZI A H.Foreign matter identification in piper nigrum samples[C]//Proceedings of the 7th International Colloquium on Signal Processing and its Applications.Washington D.C., USA:IEEE Press, 2011:1197-1204.
[8] LIANG H G, ZUO C, WEI W M.Detection and evaluation method of transmission line defects based on deep learning[J].IEEE Access, 2020, 8:38448-38458.
[9] JING J, ZHUO D, ZHANG H, et al.Fabric defect detection using the improved YOLOv3 model[J].Journal of Engineered Fibers and Fabrics, 2020, 15(1):1-10.
[10] CAO Z Y, LI X R, ZHAO L Y.Object detection in VHR image using transfer learning with deformable convolution[C]//Proceedings of 2019 IEEE International Geoscience and Remote Sensing Symposium.Washington D.C., USA:IEEE Press, 2019:326-329.
[11] LI M, HSU W, XIE X D, et al.SACNN:self-attention convolutional neural network for low-dose CT denoising with self-supervised perceptual loss network[J].IEEE Transactions on Medical Imaging, 2020, 7:2289-2301.
[12] LIU Q M, JIA R S, ZHAO C Y, et al.Face super-resolution reconstruction based on self-attention residual network[J] IEEE Access, 2020, 8:4110-4121.
[13] ZHANG H, GOODFELLOW I, METAXAS D, et al.Self-attention generative adversarial network[EB/OL].[2021-01-03].https://arxiv.org/abs/1805.08318.
[14] BA J L, MNIH V, KAVUKCUOGLU K.Multiple object recognition with visual attention[EB/OL].[2021-01-03].https://arxiv.org/abs/1412.
[15] REN D W, ZUO W M, HU Q H, et al.Progressive image deraining networks:a better and simpler baseline[C]//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2019:1926-1932.
[16] VASWANI A, SHAZEER N, PARMAR N, et al.Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems.Washington D.C., USA:IEEE Press, 2017:5998-6008.
[17] SANDLER M, HOWARD A, ZHU M L, et al.Mobilenetv2:inverted residuals and linear bottlenecks[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2018:1793-1800.
[18] DONAHUE J, HENDRICKS L A, ROHRBACH M, et al.Long-term recurrent convolutional networks for visual recognition anddescription[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39:677-691.
[19] FU X Y, HUANG J B, ZENG D L, et al.Removing rain from single images via a deep detail network[C]//Proceedings of 2017 IEEE International Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:1715-1723.
[20] YANG W H, ROBBY T, FENG J S, et al.Deep joint rain detection and removal from a single image[C]//Proceedings of 2017 IEEE International Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:1357-1366.
[21] LI X, WU J L, LIN Z C, et al.Recurrent squeeze-and-excitation context aggregation net for single image deraining[C]//Proceedings of 2018 European Conference on Computer Vision.Berlin, Germany:Springer, 2018:262-277.

选择文件类型/文献管理软件名称

选择包含的内容

基于刀具刃口显微图像的附着物去除网络

Attachment Removal Network Based on Micro Image of Tool Edge

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	刘豪, 吴红兰, 房宇轩. 结合全局上下文信息的高效人体姿态估计[J]. 计算机工程, 2023, 49(7): 102-109.
[2]	沈学利, 韩倩雯. 基于注意力机制的场感知点击率预测模型[J]. 计算机工程, 2023, 49(3): 80-86,94.
[3]	刘杭, 殷歆, 陈杰, 罗恒. 基于混合网络模型的多维时间序列预测[J]. 计算机工程, 2023, 49(1): 121-129.
[4]	郑秋梅, 徐林康, 王风华, 林超. 基于改进自注意力机制的金字塔场景解析网络[J]. 计算机工程, 2023, 49(1): 242-249.
[5]	赵国川, 王姮, 张华, 庞杰, 周建. 基于完全自注意力的水电枢纽缺陷识别方法[J]. 计算机工程, 2022, 48(9): 277-285.
[6]	吴迪, 王梓宇, 赵伟超. ELMo-CNN-BiGRU双通道文本情感分类模型[J]. 计算机工程, 2022, 48(8): 105-112.
[7]	陈可嘉, 刘惠. 基于改进BiGRU-CNN的中文文本分类方法[J]. 计算机工程, 2022, 48(5): 59-66,73.
[8]	韩磊, 高永彬, 史志才. 基于稀疏Transformer的雷达点云三维目标检测[J]. 计算机工程, 2022, 48(11): 104-110,144.
[9]	史占堂, 马玉鹏, 赵凡, 马博. 基于CNN‐Head Transformer编码器的中文命名实体识别[J]. 计算机工程, 2022, 48(10): 73-80.
[10]	赵宏, 郭岚, 陈志文, 郑厚泽. 基于多模态融合与多层注意力的视频内容文本表述研究[J]. 计算机工程, 2022, 48(10): 45-54.
[11]	黄彬, 胡立坤, 张宇. 基于自适应权重的改进Census立体匹配算法[J]. 计算机工程, 2021, 47(5): 189-196.
[12]	马科伟, 张锲石, 康宇航, 任子良, 程俊. 移动机器人中视觉里程计技术综述[J]. 计算机工程, 2021, 47(11): 1-10.
[13]	赵亚南, 刘渊, 宋设. 融合多头自注意力机制的金融新闻极性分析[J]. 计算机工程, 2020, 46(8): 85-92.
[14]	段大高, 梁少虎, 赵振东, 韩忠明. 基于自注意力机制的中文标点符号预测模型[J]. 计算机工程, 2020, 46(5): 291-297.
[15]	官洪运, 欧阳江坤, 杨益伟, 吴炜. 基于特征融合的改进LCT跟踪算法[J]. 计算机工程, 2019, 45(8): 241-247.

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于刀具刃口显微图像的附着物去除网络

Attachment Removal Network Based on Micro Image of Tool Edge

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献

相关文章 15

编辑推荐

Metrics

本文评价