计算机科学 ›› 2026, Vol. 53 ›› Issue (6A): 250300143-7.doi: 10.11896/jsjkx.250300143
楚春雨1, 姜飞龙2
CHU Chunyu1, JIANG Feilong2
摘要： 现有基于深度学习的水表读数识别方法一般是孤立地对水表的每一位数字或指针进行识别,再将各个位的识别结果进行简单拼接从而得到最终结果。然而,由于水表计数齿轮之间存在咬合间隙、水表自身可能存在结构误差以及拍摄角度等原因,可能会出现水表字轮数字显示不完整、字轮转到两个数字之间、指针指示偏差等情况,此时若简单地将每一位数字或指针的识别结果进行组合就会导致最终识别结果的错误。针对上述问题,文中提出一种基于深度学习与先验校正的水表读数识别方法,该方法基于PaddlePaddle框架,使用轻量化模型架构MobileNetV3与SVTR对字轮区域进行读取,同时使用图像处理技术读取指针读数,最后充分利用水表字轮各位数字以及各位指针读数之间的关联性先验知识对识别结果进行校正。文中讨论了字轮区域与指针区域的识别与校正方法,将其应用在水表图片进行实验测试,并与已有方法进行对比,结果表明文中提出的方法能够有效提高水表读数识别结果的准确性。
中图分类号:
| [1] KANG X,SUN X G,WAN L.Detection and recognition of water meter readings in complex scenes [J].Journal of Computer Applications,2019,39(S2):63-67. [2] JI X S,TAN K K.Water meter reading recognition methodbased on convolutional neural network [J].Transducer and Microsystem Technologies,2020,39(12):130-133. [3] YANG F,JIN L,LAI S,et al.Fully convolutional sequence recognition network for water meter number reading[J].IEEE Access,2019,7:11679-11687. [4] QI X L.Research and Development of Remote Water MeterReading Recognition Technology Based on Deep Learning [D].Zhenjiang:Jiangsu University of Science and Technology,2023. [5] WU Y C.Research on Automatic Recognition Algorithm forWater Meter Readings [D].Shenyang:Shenyang University of Technology,2024. [6] WANG Y,XIANG X.GMS-YOLO:an enhanced algorithm for water meter reading recognition in complex environments[J].Journal of Real-Time Image Processing,2024,21(5):173. [7] WAN J L,WANG H F,GUAN M Y,et al.Automatic recognition method for pointer-type instrument readings in substations based on Faster R-CNN and U-Net [J].Power System Techno-logy,2020,44(8):3097-3105. [8] CHEN L,WU X,SUN C,et al.An intelligent vision recognitionmethod based on deep learning for pointer meters[J].Measurement Science and Technology,2023,34(5):055410. [9] JI D,ZHANG W,ZHAO Q,et al.Correction and pointer reading recognition of circular pointer meter[J].Measurement Science and Technology,2022,34(2):025406. [10] ZHOU D,YANG Y,ZHU J,et al.Intelligent reading recogni-tion method of a pointer meter based on deep learning in a real environment[J].Measurement Science and Technology,2022,33(5):055021. [11] Jianyang L,Junrong B,Bingjin L,et al.The Character Recognition Method Based on OCR[C]//2023 26th ACIS International Winter Conference on Software Engineering,Artificial Intelligence,Networking and Parallel/Distributed Computing(SNPD-Winter).IEEE,2023:92-95. [12] DU Y,LI C,GUO R,et al.Pp-ocr:A practical ultra lightweight ocr system[J].arXiv:2009.09941,2020. [13] HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2016:770-778. [14] LI C,LIU W,GUO R,et al.PP-OCRv3:More attempts for the improvement of ultra lightweight OCR system[J].arXiv:2206.03001,2022. [15] DU Y,CHEN Z,JIA C,et al.Svtr:Scene text recognition with a single visual model[J].arXiv:2205.00159,2022. [16] LIU J,BAO J,LI B,et al.The Character Recognition Method Based on OCR[C]//2023 26th ACIS International Winter Conference on Software Engineering, Artificial Intelligence,Networking and Parallel/Distributed Computing(SNPD-Winter).IEEE,2023:92-95. [17] KOKKU A,CHAKRAVARTHY S.A complete OCR systemfor Tamil magazine documents[J].Guide to OCR for Indic Scripts:Document Recognition and Retrieval,2010:147-162. [18] RONG W,LI Z,ZHANG W,et al.An improved CANNY edge detection algorithm[C]//2014 IEEE International Conference on Mechatronics and Automation.IEEE,2014:577-582. [19] ZHANG C,SHI L,ZHANG D,et al.Pointer meter recognition method based on Yolov7 and hough transform[J].Applied Sciences,2023,13(15):8722. |
|
||
首页