传感器世界2024,Vol.30Issue(8) :35-40.DOI:10.16204/j.sw.issn.1006-883X.2024.08.006

基于传感器阵列的SF6仪表图像识别方法

The SF6 Instrument Image Recognition Method Based on the Sensor Array

赵炜 饶雪梅 谢榕 李杰 刘鑫 李汉豪
传感器世界2024,Vol.30Issue(8) :35-40.DOI:10.16204/j.sw.issn.1006-883X.2024.08.006

基于传感器阵列的SF6仪表图像识别方法

The SF6 Instrument Image Recognition Method Based on the Sensor Array

赵炜 1饶雪梅 1谢榕 1李杰 2刘鑫 2李汉豪2
扫码查看

作者信息

  • 1. 贵州电网有限责任公司毕节供电局,贵州毕节 563200
  • 2. 广州安电测控技术有限公司,广东广州 510000
  • 折叠

摘要

SF6仪表图像中存在多个亮暗区域、反光和阴影等问题,以单一传感器为基础的识别方法缺陷明显.为此,引入传感器阵列,提出SF6仪表图像识别方法.将索尼的IMX219作为核心图像传感器,设计传感器阵列,全方位捕捉SF6仪表图像信息,在TCP/IP协议下实时处理和传输传感器阵列采集的数据.通过特征向量和相似度参数对图像去噪与多尺度小波变换,得到不同尺度下的纹理特征.利用支持向量机(Support Vector Machine,SVM)实现对SF6仪表图像信息的识别.测试结果表明,该设计图像识别方法不仅具有优越的去噪效果,对于具体示数的识别结果也有更低的误差.

Abstract

SF6 instrument image light dark area,there are multiple reflections and shadows,defect recognition method based on single sensor is obvious. Therefore,the sensor array is introduced and the image recognition method of SF6 instrument is proposed. Using SONY's IMX219 as the core image sensor,the sensor array is designed to capture the SF6 instrument image information in all aspects,and the data collected by the sensor array is processed and transmitted in real time under TCP/IP protocol. Image denoising and multi-scale wavelet transform are used to obtain the texture features at different scales through feature vectors and similarity parameters. Support vector machine (SVM) recognition was used to recognize SF6 instrument image information. In the test results,the image recognition method designed in this paper not only has superior denoising effect,but also has lower error for the recognition results of specific symbols.

关键词

传感器阵列/SF6仪表/信息识别/索尼IMX219/图像传感器/图像去噪

Key words

sensor array/SF6 instrument/information recognition/SONY IMX219/image sensor/image denoising

引用本文复制引用

基金项目

2023年指针式(数字式)表计图像数据智能识别研究项目(GZDW-FW-0490-22-01)

出版年

2024
传感器世界
北京信息科技大学

传感器世界

影响因子:0.196
ISSN:1006-883X
段落导航相关论文