期刊文献+

数字电表图像的检测与识别 被引量:3

Detection and recognition of digital ammeter image
下载PDF
导出
摘要 为降低传统机械式电表读取的人工成本,提升电子屏图像数据的提取效率,文中设计一种基于图像处理的电表读数智能检测系统。该系统主要由数字检测与数字识别两个模块构成。数字检测模块通过特有图像特征对电子屏区域进行定位与提取,随后对区域内的数字由粗至精逐步进行精细框选,实现对单个数字的分割过程。数字识别模块采用大量数字图像对K最近邻分类算法进行训练,得到识别模型后,判断检测模块中分割数字的类别。以摄像头拍摄的特定电表为检测对象,经过工厂实地图像取样,通过编程和实验结果分析,证明文中检测系统能够较好地检测出电子屏幕中显示的数字个数并准确判断出数字类别,识别度高于97%。此电表读数智能检测系统可有效节省人力投入,降低出错率,能够依据实际需求对模型进行调整,从而应用于不同类型的电表图像识别。 In order to reduce the labor cost of traditional mechanical meter reading and improve the efficiency of electronic screen image data extraction,an intelligent meter reading detection system based on image processing is designed.The system is mainly composed of two modules:digital detection and digital recognition.The digital detection module is used to locate and extract the electronic screen area by means of the unique image features,and then select the numbers in the area from coarse to fine to realize the segmentation process of single number.The digital recognition module can be used to train the K⁃nearest neighbor model with a large number of digital images,which can effectively determine the category of the segmented numbers.The system can take the specific ammeter captured by the camera as the detection object,sample the images in the factory,and realize the algorithm by code.The system can detect the number of numbers displayed on the electronic screen and recognize the number category accurately,and the recognition rate is higher than 97%.This intelligent ammeter reading detection system can effectively save manpower input,reduce the error rate,and it can be applied to different types of ammeter image recognition by adjusting the model according to the actual needs.
作者 沈美丽 SHEN Meili(School of Science,Qingdao University of Technology,Qingdao 266520,China)
出处 《现代电子技术》 2022年第16期110-114,共5页 Modern Electronics Technique
关键词 数字提取 ROI提取 数字识别 K最近邻分类 图像分割 数字电表 图像特征 数字切割 欧氏距离 number extraction ROI extraction number recognition K⁃nearest neighbor classifier image segmentation digital ammeter image features digital cutting Euclidean distance
  • 相关文献

参考文献13

二级参考文献110

共引文献87

同被引文献30

引证文献3

二级引证文献10

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部