Research on Fuse Coil Character Segmentation and Recognition Technology

doi:10.3778/j.issn.1002-8331.2105-0499

Abstract

Abstract: In view of the fact that the coil characters of fuse are exposed excessively and contain a lot of noise during the collection process and the coil characters are compressed, the projection method combined with the prior knowledge of characters is used to segment the characters, and the expanded clear characters are used for LeNet network recognition. In recognition, each character in the coil is firstly marked and trained, and the improved Faster R-CNN positioning is used for recognition. The improved parts include replacing Backbone, finding the best Anchor parameters, replacing ROI Pooling with ROI Align and designing loss function. Compared with other networks, the experimental result shows that the character recognition rate of the improved network is up to 99.91%, which has good robustness and generalization ability, and has certain popularization and reference value for optical character recognition.

Key words: fuse coil, projection method, deep learning, improved Faster R-CNN, positioning for recognition

摘要： 针对保险丝线圈字符在采集过程中曝光过度含有大量的噪声以及线圈字符被压缩情况，使用投影法结合字符先验知识进行字符分割，并将膨胀之后的清晰字符用于LeNet网络识别。在识别中，将线圈中每个字符进行标注训练，使用改进的Faster R-CNN定位作识别，改进的部分为替换Backbone、寻找最佳Anchor参数、将RoI Pooling替换成RoI Align和设计损失函数。对比未改进网络和其他网络，实验结果表明，改进后的网络字符识别率达99.91%，具有良好的鲁棒性和泛化能力，对光学字符识别具有一定的推广和借鉴价值。

关键词: 保险丝线圈, 投影法, 深度学习, 改进的Faster R-CNN, 定位作识别

OUYANG Jiabin, HU Weiping, LIU Beibei, LIU Wenyang. Research on Fuse Coil Character Segmentation and Recognition Technology[J]. Computer Engineering and Applications, 2022, 58(23): 300-306.

欧阳家斌, 胡维平, 刘北北, 刘文扬. 保险丝线圈字符识别技术研究[J]. 计算机工程与应用, 2022, 58(23): 300-306.

References

[1] 孙阳光，蔡志华.一种基于局部自适应的女书字符图像分割算法[J].微电子学与计算机，2014，31（3）：142-145.
SUN Y G，CAI Z H.Research on image segmentation method for NvShu characters based on local adaptive algorithm[J].Microelectronics ＆ Computer，2014，31（3）：142-145.
[2] 李旭，徐舒畅，尤玉才，等.基于聚类分析的个性化美国车牌分割算法[J].浙江大学学报（工学版），2021，46（12）：2155-2159.
LI X，XU S C，YOU Y C，et al.Segmentation method for personalized American car plate based on clustering analysis[J].Journal of Zhejiang University（Engineering Edition），2021，46（12）：2155-2159.
[3] 张艰，赵宇明，陈德权.分时电表表盘数字读数的自动识别系统[J].计算机工程，2005，31（5）：178-180.
ZHANG J，ZHAO Y M，CHEN D Q.Automatic recognition system for numeric characters on time-sharing ammeter dial plate[J].Computer Engineering，2005，31（5）：178-180.
[4] 姚可欣，曹卫群.Trans-Net：基于迁移学习的手写简笔画识别[J].计算机工程与应用，2021，57（3）：182-188.
YAO K X，CAO W Q.Trans-Net：stick figure recognition based on transfer learning[J].Computer Engineering and Applications，2021，57（3）：182-188.
[5] 王智文.基于改进BP神经网络的车牌字符识别[J].微电子学与计算机，2011，28（9）：66-69.
WANG Z W.Vehicle license plate recognition based on improved BP network[J].Microelectronics ＆Computer，2011，28（9）：66-69.
[6] LECUN Y，BOTTOU L，BENGIO Y，et al.Gradient-based learning applied to document recognition[J].Proceedings of the IEEE，1998，86（11）：2278-2324.
[7] LIAO M，SHI B，BAI X，et al.TextBoxes：a fast text detector with a single deep neural network[C]//Proceedings of AAAI，2017：4161-4167.
[8] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//European Conference on Computer Vision.Cham：Springer，2016.
[9] ZHI T，HUANG W，TONG H，et al.Detecting text in natural image with connectionist text proposal network[C]//European Conference on Computer Vision，2016.
[10] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition，2014.
[11] GIRSHICK R.Fast R-CNN[C]//IEEE International Conference on Computer Vision（ICCV），2015.
[12] UIJLINGS J R R，VAN DE SANDE K E A，GEVERS T，et al.Selective search for object recognition[J].International Journal of Computer Vision，2013，104（2）：154-171.
[13] REN S Q，HE K M，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39（6）：1137-1149.
[14] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：779-788.
[15] REDMON J，FARHADI A.Yolov3：an incremental improvement[J].arXiv：1804.02767，2018.
[16] HE K M，ZHANG X Y，REN S Q，et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recogintion，2016.
[17] 周本金，陶以政，纪斌，等.最小化误差平方和k-means初始聚类中心优化方法[J].计算机工程与应用，2018，54（15）：48-52.
ZHOU B J，TAO Y Z，JI B，et al.Optimizing k-means initial clustering centers by minimizing sum of squared error[J].Computer Engineering and Applications，2018，54（15）：48-52.
[18] 晏文仲，李光.基于字符分割与新型LENET网络的票据识别算法[J].包装工程，2020，41（21）：244-250.
YAN W Z，LI G.Ticket recognition algorithm based on character segmentation and new LENET network[J].Packaging Engineering，2020，41（21）：244-250.
[19] 刘黎，韩睿，韩译锋，等.改进的Faster-RCNN目标检测方法在变电站悬挂异物检测中的应用[J].电测与仪表，2021，58（1）：142-146.
LIU L，HAN R，HAN Y F，et al.Application of an improved Faster-RCNN object detection method in the detection of suspended foreign matters in substation[J].Electrical Measurement & Instrumentation，2021，58（1）：142-146.
[20] 訾梦超.基于深度学习的蓝宝石衬底表面缺陷检测与分类算法研究[D].泉州：华侨大学，2020.
ZI M C.Research on surface defect detection and classification algorithm of sapphire based on deep learning[D].Quanzhou：Huaqiao University，2020.
[21] 张博豪.基于深层神经网络的工业图像检测技术研究[D].北京：北方工业大学，2020.
ZHANG B H.Research on image detection technology of industrial images based on deep neural network[D].Beijing：North China University of Technology，2020.