基于U-Net网络的柱面透镜视觉定位策略

doi:10.3969/j.issn.1004-132X.2023.05.001

中国机械工程 ›› 2023, Vol. 34 ›› Issue (05): 505-514.DOI: 10.3969/j.issn.1004-132X.2023.05.001

基于U-Net网络的柱面透镜视觉定位策略

陈逢军1;吕继阳1;胡天2;梁小生3

1.湖南大学机械与运载工程学院，长沙，410082
2.湖南艾凯瑞斯智能科技有限公司，长沙，410205
3.湖南戴斯光电有限公司，长沙，410600

出版日期:2023-03-10 发布日期:2023-03-27
作者简介:陈逢军，男，1979年生，副教授、博士研究生导师。研究方向为智能制造、超精密加工与控制。E-mail:abccfj@126.com。
基金资助:
湖南省重点研发计划（2021GK2025）；国家自然科学基金（51975204）；湖南省自然科学基金（2021JJ30103）

Visual Positioning Strategy of Cylindrical Lens Based on U-Net Network

CHEN Fengjun1;LYU Jiyang1;HU Tian2;LIANG Xiaosheng3

1.College of Mechanical and Vehicle Engineering,Hunan University,Changsha,410082
2.Hunan Accuracy Intelligent Technology Co.,Ltd.,Changsha,410205
3.Hunan Dayoptronics Co.,Ltd.,Changsha,410600

Online:2023-03-10 Published:2023-03-27

摘要/Abstract

摘要： 针对柱面透镜定位在摆盘工艺过程中产品良率低、生产作业时间长的问题，研究了一种基于U-Net网络的柱面透镜视觉定位策略。该策略结合柱面透镜摆盘工艺及特点，开发了柱面透镜视觉摆盘系统。将Blob分析和模板匹配作为前期视觉定位方法，通过高精度匹配获取工件点位信息以生成分割图对采集的原图进行标注，选取损失值最优的学习率对U-Net网络进行模型训练，以实现实际工件的分割定位。经过U-Net网络与高精度模板匹配针对不同型号的工件定位对比发现，相较于高精度模板匹配，U-Net网络可实现亚像素级别的定位误差。实践应用结果表明，该柱面透镜视觉摆盘系统视觉定位精度高，具有较强的产品适应能力，产品偏心合格率可达96%以上，满足工业摆盘任务要求。

关键词: U-Net网络, 柱面透镜, 视觉定位, 摆盘系统

Abstract: Aiming at the problems of low product yield and long production operation time during the plating processes of cylindrical lens positioning, a visual positioning correction strategy of cylindrical lens was studied based on U-Net network. The cylindrical lens plating processes and characteristics were combined in the strategy to develop a cylindrical lens visual plating system. The Blob analysis and template matching were used as the early visual positioning method, and the workpiece point information was obtained through high-precision matching to generate a segmentation diagram to annotate the original map of the acquisition, and the optimal learning rate of the loss values was selected to train the U-Net network to achieve the segmentation and positioning of the actual workpiece. U-Net was compared to high-precision template matching for different models of workpiece positioning. It is found that U-Net may achieve sub-pixel level positioning error relative to high-precision template matching. Practical application results show that the visual positioning accuracy of the cylindrical lens visual plating system has high accuracy and strong product adaptability, and the eccentric pass rate of the product may reach more than 96%, which meets the requirements of industrial plating tasks.

Key words: U-Net network, cylindrical lens, visual positioning, plating system

中图分类号:

TP23

陈逢军, 吕继阳, 胡天, 梁小生. 基于U-Net网络的柱面透镜视觉定位策略[J]. 中国机械工程, 2023, 34(05): 505-514.

CHEN Fengjun, LYU Jiyang, HU Tian, LIANG Xiaosheng. Visual Positioning Strategy of Cylindrical Lens Based on U-Net Network[J]. China Mechanical Engineering, 2023, 34(05): 505-514.

参考文献

［1］OGAWA T, YAMAZAWA Y, KAWAI S, et al. A Novel Monochromator with Offset Cylindrical Lenses and Its Application to a Low-voltage Scanning Electron Microscope［J］. Microscopy and Microanalysis, 2022, 28(2):412-424.
［2］BADEEVA E, MURASHKINA T, MOTIN A, et al. Measuring Setup for Experimental Research of Two-coordinate Fiber-optic Acceleration Sensors with Cylindrical Lenses［J］. Sensors, 2022, 22(3):1125.
［3］XIAO D, WEI R, LIU S J, et al. Holographic Magnification System Based on Cylindrical Lens［J］. Optik, 2018, 174:707-714.
［4］CHEN G X, HUANG H Y, ZHAO Q, et al. Cylindrical Wave-based Off-axis Digital Holography with Long Field of View［J］. Optics Letters, 2022, 47(3):449-452.
［5］朱子俊, 刘顺, 韩冰, 等. 超声振动复合研磨K9光学玻璃工艺研究［J］. 表面技术, 2020, 49(4):74-80.
ZHU Zijun, LIU Shun, HAN Bing, et al. Study on Hybrid Technology of Ultrasonic Vibration Assisted Abrasive Lapping K9 Optical Glass ［J］. Surface Technology, 2020, 49(4):74-80.
［6］贾坡, 田建艳, 杨英波, 等. 基于机器视觉的滚抛磨块缺陷检测方法［J］. 金刚石与磨料磨具工程, 2021, 41(1):76-82.
JIA Po, TIAN Jianyan, YANG Yingbo, et al. Defect Detection Method of Abrasive Block Based on Machine Vision［J］. Diamond and Abrasives Engineering, 2021, 41(1):76-82.
［7］ZHOU A, AI B, QU P, et al. Defect Detection for Highly Reflective Rotary Surfaces:an Overview［J］. Measurement Science and Technology, 2021, 32(6):062001.
［8］弯艳玲, 崔普, 徐丽宁, 等. 基于图像处理技术界定微纳复合织构防覆冰性能［J］. 表面技术, 2019, 48(08):54-58.
WAN Yanling, CUI Pu, XU Lining, et al. Anti-icing Performance of Micro-nano Composite Texture Based on Image Processing Technology ［J］. Surface Technology, 2019, 48(08):54-58.
［9］何彦, 肖圳, 李育锋, 等. 使用CNN-SVR的汽车组合仪表组装质量预测方法［J］. 中国机械工程, 2022, 33(7):825-833.
HE Yan, XIAO Zhen, LI Yufeng, et al. An Assembly Quality Prediction Method for Automotive Instrument Clusters Using CNN-SVR［J］. China Mechanical Engineering, 2022, 33(7):825-833.
［10］朱永国, 张文博, 邓正平, 等. 基于激光跟踪仪和机器视觉的飞机翼身对接装配偏差动态综合修正［J］. 机械工程学报, 2019, 55(24):187-196.
ZHU Yongguo, ZHANG Wenbo, DENG Zhengping, et al. Dynamic Synthesis Correction of Deviation for Aircraft Wing-fuselage Docking Assembly Based on Laser Tracker and Machine Vision ［J］. Journal of Mechanical Engineering, 2019, 55(24):187-196.
［11］胡刚. 基于机器视觉技术的自动摆盘设备设计开发［D］. 杭州：浙江大学, 2016.
HU Gang. The Design and Development of Automatic Wobble Plates Equipment Based on Machine Vision Technology［D］. Hangzhou:Zhejiang University, 2016.
［12］CHEN F, YE X, YIN S, et al. Automated Vision Positioning System for Dicing Semiconductor Chips Using Improved Template Matching Method［J］. International Journal of Advanced Manufacturing Technology, 2019, 100(9/12):2669-2678.
［13］邓捷, 李伟仙, 吴思进. 基于灰度及几何特征的微晶玻璃缺陷检测［J］. 光学技术, 2021, 47(4):428-431.
DENG Jie, LI Weixian, WU Sijin. Detection of Defects in Glass-Ceramics Based on Grayscale and Geometric Features［J］. Optical Technology, 2021, 47(4):428-431.
［14］ZHONG F, HE S, LI B. Blob Analyzation-based Template Matching Algorithm for LED Chip Localization［J］. International Journal of Advanced Manufacturing Technology, 2017, 93(1/4):55-63.
［15］LECUN Y, BENGIO Y, HINTON G. Deep Learning［J］. Nature, 2015, 521(7553):436.
［16］杨易, 高骏, 谷正气, 等. 基于GA-BP的汽车风振噪声声品质预测模型［J］. 机械工程学报, 2021, 57(24):241-249.
YANG Yi, GAO Jun, GU Zhengqi, et al. Acoustic Quality Prediction Model of Automobile Wind Vibration Noise Based on GA-BP［J］. Journal of Mechanical Engineering, 2021, 57(24):241-249.
［17］张胜文, 周曦, 李滨城, 等. 基于图像深度学习的零件加工特征信息提取方法［J］. 中国机械工程, 2022, 33(3):348-355.
ZHANG Shengwen, ZHOU Xi, LI Bincheng, et al. Feature Information Extraction Method for Part Machining Based on Image Deep Learning［J］. China Mechanical Engineering, 2022, 33(3):348-355.
［18］REN S, HE K, GIRSHICK R, et al. Faster R-CNN:Towards Real-time Object Detection with Region Proposal Networks［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6):1137-1149.
［19］RONNEBERGER O, FISCHER P, BROX T. U-Net:Convolutional Networks for Biomedical Image Segmentation［C］∥International Conference on Medical Image Computing and Computer-assisted Intervention. Cham, 2015:234-241.
［20］李潇凡, 王胜强, 翁轩, 等. 基于UNet深度学习算法的东海大型漂浮藻类遥感监测［J］. 光学学报, 2021, 41(2):18-26.
LI Xiaofan, WANG Shengqiang, WENG Xuan, et al. Remote Sensing Monitoring of Large Floating Algae inthe East China Sea Based on UNet Deep Learning Algorithm［J］. Acta Optica Sinica, 2021, 41(2):18-26.
［21］李连伟,秦世引.基于轻量级U-Net深度学习的人体安检隐匿违禁物的实时检测［J］.电子与信息学报,2022,44(10):3435-3446.
LI Lianwei, QIN Shiyin. Real-time Detection of Hiding Contraband in Human Body during the Security Check Based on Lightweight U-Net with Deep Learning［J］. Journal of Electronics & Information Technology, 2022, 44(10):3435-3446.

基于U-Net网络的柱面透镜视觉定位策略

Visual Positioning Strategy of Cylindrical Lens Based on U-Net Network

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

编辑推荐

Metrics