Muhammad Waqar, M. Waris, Esha Rashid, Nudrat Nida, Shah Nawaz, M. Yousaf
{"title":"Meter Digit Recognition Via Faster R-CNN","authors":"Muhammad Waqar, M. Waris, Esha Rashid, Nudrat Nida, Shah Nawaz, M. Yousaf","doi":"10.1109/ICRAI47710.2019.8967357","DOIUrl":null,"url":null,"abstract":"The current method of meter reading is manual and error-prone in developing countries. A meter reader logs the reading to calculate the cost of electricity. In recent years, there have been multiple efforts to provide automated solutions to read the meter digits. However, the existing systems extract reading based on a specific meter topology. In this paper, we propose an approach based on Faster R-CNN to extract and recognize digits in an electric meter. We compared our method against several state-of-the-art object detection methods. The proposed approach is robust against different lightening conditions, severe perspective distortions and blurred images. In addition, it is scaleinvariant. Furthermore, we created a new dataset consisting of 10310 images taken from electricity companies in Pakistan to benchmark meter digit recognition task. Experimental results shows the high accuracy of the proposed approach on the created electricity meter dataset.","PeriodicalId":429384,"journal":{"name":"2019 International Conference on Robotics and Automation in Industry (ICRAI)","volume":"71 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 International Conference on Robotics and Automation in Industry (ICRAI)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICRAI47710.2019.8967357","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 7
Abstract
The current method of meter reading is manual and error-prone in developing countries. A meter reader logs the reading to calculate the cost of electricity. In recent years, there have been multiple efforts to provide automated solutions to read the meter digits. However, the existing systems extract reading based on a specific meter topology. In this paper, we propose an approach based on Faster R-CNN to extract and recognize digits in an electric meter. We compared our method against several state-of-the-art object detection methods. The proposed approach is robust against different lightening conditions, severe perspective distortions and blurred images. In addition, it is scaleinvariant. Furthermore, we created a new dataset consisting of 10310 images taken from electricity companies in Pakistan to benchmark meter digit recognition task. Experimental results shows the high accuracy of the proposed approach on the created electricity meter dataset.