|本期目录/Table of Contents|

[1]单文轩,徐永祥,白福忠,等.基于改进YOLOv3的工业指针式仪表检测[J].工业仪表与自动化装置,2023,(04):64-70.[doi:10.19950/j.cnki.cn61-1121/th.2023.04.012]
　SHAN Wenxuan,XU Yongxiang,BAI Fuzhong,et al.Industrial pointer meter detection based on improved YOLOv3[J].Industrial Instrumentation & Automation,2023,(04):64-70.[doi:10.19950/j.cnki.cn61-1121/th.2023.04.012]
点击复制

基于改进YOLOv3的工业指针式仪表检测

《工业仪表与自动化装置》[ISSN:1000-0682/CN:61-1121/TH]

卷:
期数:
2023年04期
页码:
64-70
栏目:
出版日期:
2023-08-15

文章信息/Info

Title:
Industrial pointer meter detection based on improved YOLOv3
文章编号:
1000-0682（2023）03-0064-07
作者:
单文轩1; 徐永祥2; 白福忠2; 高晓娟3
1.内蒙古工业大学 机械工程学院；
2.内蒙古自治区特殊服役智能机器人重点实验室；
3.内蒙古工业大学 航空学院，内蒙古 呼和浩特 010051
Author(s):
SHAN Wenxuan1;  XU Yongxiang2;  BAI Fuzhong2;  GAO Xiaojuan3
1.School of mechanical engineering, Inner Mongolia University of Technology;
2. Inner Mongolia Key Laboratory of Special Service Intelligent Robotics;
3.School of Astronautics, Inner Mongolia University of Technology, Inner Mongolia Huhhot 010051, China
关键词:
YOLOv3; 指针式仪表检测; Kmeans++; 注意力机制
Keywords:
YOLOv3;  pointer meter detection;  Kmeans++;  attention mechanism
分类号:
TP391.41
DOI:
10.19950/j.cnki.cn61-1121/th.2023.04.012
文献标志码:
A
摘要:
为解决现有工业指针式仪表检测存在识别精度低、检测信息缺失、检测速度慢、成本较高的问题，提出一种改进的YOLOv3检测算法。新算法通过Kmeans++聚类获得先验框尺寸，选用轻量级网络Darknet-19 作为主干网络并减少预测层数，引入注意力机制同时调整样本损失函数，达到损失函数快速收敛的效果。消融实验结果表明，改进的YOLOv3算法对工业指针式仪表检测精度达98.16%，检测速度相比原版YOLOv3网络提升一倍，检测结果信息完整，训练消耗资源降低3倍。算法在鲁棒性、实时性、实用性方面优势明显。
Abstract:
To solve the problems of low recognition accuracy, lack of detection information, slow detection speed and high cost in the existing detection methods of industrial pointer meters, an improved YOLOv3 detection algorithm is proposed. The new algorithm obtains the size of priori frame through Kmeans++ clustering, selects the lightweight network Darknet-19 as the backbone network and reduces the number of prediction layers, introduces the attention mechanism and adjusts the sample loss function. The experimental results show that the improved YOLOv3 algorithm achieves the detection accuracy of 98.16% for industrial pointer meter and the detection speed is twice as fast as the original YOLOv3 network. In addition, the detection result information is more complete and it reduces training resource consumption by 3 times. The improved algorithm has obvious advantages in robustness, real time and practicability.

参考文献/References:

[1] 王明吉,刘博,陈秋梦,等.基于yolov3的车牌定位识别系统[J].工业仪表与自动化装置,2022(01):97-100.

[2] 郑昌庭,王俊,郑克.基于图像识别的变电站巡检机器人仪表识别研究[J].工业仪表与自动化装置,2020(05):57-61.
[3] 陈梦迟,黄文君,张阳阳,等.基于机器视觉的工业仪表识别技术研究[J].控制工程,2020,27(11):1995-2001.
[4] 卢翔,苏杨,余萱,等.基于深度学习的机房人物重识别研究[J].工业仪表与自动化装置,2021(02):104-107.
[5] ZHAO C, FAN C. The center of the circle fitting optimization algorithm based on the hough transform for crane[J]. Applied Sciences.2022.12(20): 10341-10341.
[6] 崔建国,孙长库,李玉鹏,等.基于SURF的快速图像匹配改进算法[J].仪器仪表学报,2022,43 (8):47-53.
[7].ZHOU L, YANG L, FU D, et al. SIFT-Flow-based virtual sample generation for single-sample finger vein recognition[J].Electronics.2022.11(20):3382-3382.
[8].GUO Z, TIAN Y, MAO W. A robust Faster R-CNN model with feature enhancement for rust detection of transmission line fitting[J]. Sensors.2022.22(20): 7961-7961.
[9] YAO J, WANG Z, LIU C, et al. Detection method of crushing mouth loose material blockage based on SSD algorithm[J]. Sustainability.2022.14(21): 14386-14386.
[10] 万伟彤,李长峰,朱华波,等.轻量化CenterNet网络的二维条码定位算法[J].电子测量与仪器学报,2022,36(05):128-135.
[11].段禄成,谭保华,余星雨.基于改进YOLOv3的酒瓶盖瑕疵检测算法[J].电子测量技术,2022,45(15):130-137.
[12] WANG L, YANG S, SHAN Y , et al. Automatic thyroid nodule recognition and diagnosis in ultrasound imaging with the YOLOv2 neural network.[J]. World journal of surgical oncology. 2019.17(1): 12-12.
[13] 郭宇强,易映萍.变电站指针式仪表示数识别方法研究[J].软件导刊,2022,21(03):62-66.
[14].刘家乐,吴怀宇,陈志环.改进YOLOv3的工业指针式仪表检测方法[J].计算机工程与设计,2021,42(7): 2001-2008.
[15] 赵辉,姜立锋,王红君,等.基于机器视觉的指针式仪表检测[J].科学技术与工程,2021,21(34):14665-14672.

相似文献/References:

[1]王明吉a,刘 博a,陈秋梦a,等.基于yolov3的车牌定位识别系统[J].工业仪表与自动化装置,2022,(01):97.[doi:10.19950/j.cnki.cn61-1121/th.2022.01.020]
　WANG Mingjia,LIU Boa,CHEN Qiumenga,et al.License plate location and recognition system based on yolov3[J].Industrial Instrumentation & Automation,2022,(04):97.[doi:10.19950/j.cnki.cn61-1121/th.2022.01.020]

备注/Memo

备注/Memo:
收稿日期：2023-03-29

基金项目：
内蒙古自治区自然科学基金（2022MS0601）；
内蒙古自治区科技计划项目（2021GG0263）；
内蒙古工业大学科学研究项目（ZZ201902）

第一作者：
单文轩（1999—），男，硕士研究生，主要研究方向为深度学习、图像处理技术。E-mail:1612467069@qq.com
更新日期/Last Update: 1900-01-01