基于MobileNetV2的圆形指针式仪表识别系统

doi:10.11772/j.issn.1001-9081.2020060765

计算机应用 ›› 2021, Vol. 41 ›› Issue (4): 1214-1220.DOI: 10.11772/j.issn.1001-9081.2020060765

所属专题：前沿与综合应用

基于MobileNetV2的圆形指针式仪表识别系统

李慧慧^1,2, 闫坤^1,2, 张李轩^1,2, 刘威^1,2, 李执^1,2

1. 桂林电子科技大学信息与通信学院, 广西桂林 541004;
2. 卫星导航定位与位置服务国家地方联合工程研究中心(桂林电子科技大学), 广西桂林 541004

收稿日期:2020-06-08 修回日期:2020-11-17 出版日期:2021-04-10 发布日期:2020-12-22
通讯作者: 闫坤
作者简介:李慧慧（1993—），女，河南周口人，硕士研究生，主要研究方向：计算机视觉、机器学习、工业检测与识别；闫坤（1983—），女，陕西咸阳人，副教授，博士，主要研究方向：自适应信号处理、室内定位、5G通信、机器学习、图像处理；张李轩（1995—），男，湖南郴州人，硕士研究生，主要研究方向：计算机视觉、图像识别；刘威（1992—），男，安徽淮北人，硕士研究生，主要研究方向：机器学习、图像处理、信号处理；李执（1995—），男，广东揭西人，硕士研究生，主要研究方向：机器学习、图像处理、信号处理。
基金资助:
广西自然科学基金资助项目（2019JJA170069）。

Circular pointer instrument recognition system based on MobileNetV2

LI Huihui^1,2, YAN Kun^1,2, ZHANG Lixuan^1,2, LIU Wei^1,2, LI Zhi^1,2

1. School of Information and Communication, Guilin University of Electronic Technology, Guilin Guangxi 541004, China;
2. State and Local Joint Engineering Research Center for Satellite Navigation and Location Service;(Guilin University of Electronic Technology), Guilin Guangxi 541004, China

Received:2020-06-08 Revised:2020-11-17 Online:2021-04-10 Published:2020-12-22
Supported by:
This work is partially supported by the Natural Science Foundation of Guangxi (2019JJA170069).

摘要/Abstract

摘要： 针对目前指针式仪表识别任务在使用深度学习算法时存在模型参数量大、计算量大、准确率较低的问题，提出一种基于改进预训练MobileNetV2网络模型与圆形Hough变换相结合的圆形指针式仪表智能检测和识别系统。首先，采用Hough变换解决复杂场景内非圆形区域的干扰问题；然后，提取圆形区域以构建数据集；最后，使用基于改进预训练MobileNetV2网络模型对圆形指针式仪表进行识别。为客观反映所提模型的性能优劣，采用平均混淆矩阵来衡量模型性能。实验结果表明，该系统在圆形指针式仪表识别任务中的识别率达到99.76%。同时，将所提模型与其他5种不同的网络模型进行对比的结果表明，该模型与ResNet50的准确率最高，但在模型参数量和模型计算量方面，所提网络模型相较于ResNet50分别降低了90.51%和92.40%，可见该模型有助于进一步在移动端或嵌入式设备中部署和实现工业级的实时圆形指针式仪表检测和识别。

关键词: 圆形指针式仪表, 圆形Hough变换, 预训练模型, MobileNetV2, 平均混淆矩阵

Abstract: Aiming at the problems of large number of model parameters, large computational cost and low accuracy when using deep learning algorithms for pointer instrument recognition task, an intelligent detection and recognition system of circular pointer instrument based on the combination of improved pre-trained MobileNetV2 network model and circular Hough transform was proposed. Firstly, the Hough transform was used to solve the interference problem of non-circular areas in complex scene. Then, the circular areas were extracted to construct datasets. Finally, the circular pointer instrument recognition was realized by using the improved pre-trained MobileNetV2 network model. The average confusion matrix was used to measure the performance of the proposed model. Experimental results show that, the recognition rate of the proposed system in the recognition task of circular pointer instruments reaches 99.76%. At the same time, the results of comparing the proposed model with other five different network models show that the proposed model and ResNet50 both have the highest accuracy, but compared with ResNet50, the proposed network model has the model parameter number and model computational cost reduced by 90.51% and 92.40% respectively, verifying that the proposed model is helpful for the further deployment and implementation of industrial grade real-time circular pointer instrument detection and recognition in mobile terminals or embedded devices.

Key words: circular pointer instrument, circular Hough transform, pre-trained model, MobileNetV2, average confusion matrix

中图分类号:

TP391.4

李慧慧, 闫坤, 张李轩, 刘威, 李执. 基于MobileNetV2的圆形指针式仪表识别系统[J]. 计算机应用, 2021, 41(4): 1214-1220.

LI Huihui, YAN Kun, ZHANG Lixuan, LIU Wei, LI Zhi. Circular pointer instrument recognition system based on MobileNetV2[J]. Journal of Computer Applications, 2021, 41(4): 1214-1220.

参考文献

[1] ZHENG W,YIN H,WANG A,et al. Development of an automatic reading method and software for pointer instruments[C]//Proceedings of the 1st International Conference on Electronics Instrumentation and Information Systems. Piscataway:IEEE, 2017:1-6.
[2] 韩绍超, 徐遵义, 尹中川, 等. 指针式仪表自动读数识别技术的研究现状与发展[J]. 计算机科学,2018,45(6A):54-57.(HAN S C,XU Z Y,YIN Z C,et al. Research review and development for automatic reading recognition technology of point instruments[J]. Computer Science,2018,45(6A):54-57.)
[3] 李祖胜. 机械式仪表自动读数识别的研究[D]. 荆州:长江大学, 2018:9.(LI Z S. Research on automatic reading recognition of mechanical instruments[D]. Jingzhou:Yangtze University, 2018:9.)
[4] 朱海霞. 基于改进Hough变换和BP网络的指针仪表识别[J]. 电测与仪表,2015,52(5):11-14.(ZHU H X. Pointer instrument recognition based on BP network and improved Hough transform[J]. Electrical Measurement and Instrumentation,2015,52(5):11-14.)
[5] MAITRA S. Smart energy meter using power factor meter and instrument transformer[J]. Communications on Applied Electronics,2016,4(1):31-37.
[6] KUCHERUK V,KURYTNIK I,KULAKOV P,et al. Definition of dynamic characteristics of pointer measuring devices on the basis of automatic indications determination[J]. Archives of Control Sciences,2018,28(3):401-408.
[7] ALEGRIA F C,SERRA A C. Automatic calibration of analog and digital measuring instruments using computer vision[J]. IEEE Transaction on Instrumentation and Measurement,2000,49(1):94-99.
[8] VÁZQUEZ-FERNÁNDEZ E, DACAL-NIETO A, GONZÁLEZJORGE H,et al. A machine vision system for the calibration of digital thermometers[J]. Measurement Science and Technology, 2009,20(6):No. 065106.
[9] 孙婷, 马磊. 巡检机器人中指针式仪表示数的自动识别方法[J]计算机应用,2019,39(1):287-291.(SUN T,MA L. Automatic recognition method of pointer meter for inspection robots[J]. Journal of Computer Applications,2019,39(1):287-291.)
[10] 张雪飞, 黄山. 多类指针式仪表识别读数算法研究[J]. 电测与仪表,2020,57(16):147-152.(ZHANG X F,HUANG S. Research on pointer identifying and number reading algorithm of multi-class pointer instruments[J]. Electrical Measurement and Instrumentation,2020,57(16):147-152.)
[11] 盛庆华, 李竹, 邵展坚, 等. 基于双重霍夫空间投票的指针表自动读数方法[J]. 仪器仪表学报,2019,40(5):230-239. (SHENG Q H,LI Z,SHAO Z J,et al. Automatic reading method of pointer meter based on double Hough space voting[J]. Chinese Journal of Scientific Instrument,2019,40(5):230-239.)
[12] 万吉林, 王慧芳, 管敏渊, 等. 基于Faster R-CNN和U-Net的变电站指针式仪表读数自动识别方法[J]. 电网技术,2020,44(8):3097-3105.(WAN J L,WANG H F,GUAN M Y,et al. An automatic identification method for the reading of substation pointer-type meters using faster R-CNN and U-Net[J]. Power System Technology,2020,44(8):3097-3105.)
[13] 周杨浩, 刘一帆, 李瑮. 一种自动读取指针式仪表读数的方法[J]. 山东大学学报(工学版),2019,49(4):1-7.(ZHOU Y H, LIU Y F,LI L. An automatic reading method for pointer meter[J]. Journal of Shandong University (Engineering Science), 2019,49(4):1-7.)
[14] 钟志伟, 张寿明. 基于概率Hough变换的指针仪表去影算法[J]. 传感器与微系统,2018,37(6):126-128.(ZHONG Z W, ZHANG S M. Algorithm of pointer instrument shadow elimination based on probabilistic Hough transform[J]. Transducer and Microsystem Technologies,2018,37(6):126-128.)
[15] SANDLER M, HOWARD A, ZHU M, et al. MobileNetV2:inverted residuals and linear bottlenecks[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2018:4510-4520.
[16] 张帅. 基于中点画圆的圆检测[D]. 大连:大连理工大学, 2013:4. (ZHANG S. Circle detection by midpoint circle validation[D]. Dalian:Dalian University of Technology, 2013:4)
[17] LIU X,YU J,WANG Y,et al. Automatic localization of the fetal cerebellum on 3D ultrasound volumes[J]. Medical Physics, 2013,40(11):No. 112902.
[18] XU Z,BAO C. Detection for deformed and sheltered circular traffic signs[J]. MATEC Web of Conferences, 2015, 22:No. 03021.
[19] DJEKOUNE A O,MESSAOUDI K,AMARA K. Incremental circle Hough transform:an improved method for circle detection[J]. Optik,2017,133:17-31.
[20] HOWARD A G,ZHU M,CHEN B,et al. MobileNets:efficient convolutional neural networks for mobile vision applications[EB/OL].[2020-05-21]. https://arxiv.org/pdf/1704.04861.pdf.
[21] LI Y,ZHANG D,LEE D J. IIRNet:a lightweight deep neural network using intensely inverted residuals for image recognition[J]. Image and Vision Computing,2019,92:No. 103819.
[22] PAN S J,YANG Q. A survey on transfer learning[J]. IEEE Transactions on Knowledge and Data Engineering,2010,22(10):1345-1359.
[23] DENG X,LIU Q,DENG Y,et al. An improved method to construct basic probability assignment based on the confusion matrix for classification problem[J]. Information Sciences,2016, 340/341:250-261.

基于MobileNetV2的圆形指针式仪表识别系统

Circular pointer instrument recognition system based on MobileNetV2

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 4

编辑推荐

Metrics

[1]	刘睿珩, 叶霞, 岳增营. 面向自然语言处理任务的预训练模型综述[J]. 计算机应用, 2021, 41(5): 1236-1246.
[2]	罗俊, 陈黎飞. 基于BERT的不完全数据情感分类[J]. 计算机应用, 2021, 41(1): 139-144.
[3]	邓雄, 王洪春. 基于深度学习和特征融合的人脸活体检测算法[J]. 计算机应用, 2020, 40(4): 1009-1015.
[4]	王昆, 郑毅, 方书雅, 刘守印. 基于文本筛选和改进BERT的长文本方面级情感分析[J]. 计算机应用, 2020, 40(10): 2838-2844.