Text Detection of Transformer Based on Deep Learning Algorithm
Transformers are important equipment in the power system. At present, the text information collection of transformer nameplates is through manual, which is inefficient. Therefore, it is necessary to find a high-precision automatic detection method of transformer text information. However, the curren...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Faculty of Mechanical Engineering in Slavonski Brod, Faculty of Electrical Engineering in Osijek, Faculty of Civil Engineering in Osijek
2022-01-01
|
| Series: | Tehnički Vjesnik |
| Subjects: | |
| Online Access: | https://hrcak.srce.hr/file/398877 |
| _version_ | 1797206729425944576 |
|---|---|
| author | Yu Cheng Yiru Wan Yingjie Sima Yinmei Zhang Sanying Hu Shu Wu |
| author_facet | Yu Cheng Yiru Wan Yingjie Sima Yinmei Zhang Sanying Hu Shu Wu |
| author_sort | Yu Cheng |
| collection | DOAJ |
| description | Transformers are important equipment in the power system. At present, the text information collection of transformer nameplates is through manual, which is inefficient. Therefore, it is necessary to find a high-precision automatic detection method of transformer text information. However, the current text detection algorithms have limited ability to detect special characters on the transformer. And they will also have the problem of incomplete detection in detecting the dense text and long text on the transformer nameplate. We propose a text detection network based on segmentation to automatically calibrate the text box of transformer nameplates. Our network is based on DB (differential binarization) network. It has a new feature fusion structure, which refers to the feature fusion structure of the u-net network. The proposed network has achieved better performance than the advanced scene text detection algorithms (DB, East) on the English scene text dataset icdar2015 and the Chinese-English mixed scene text dataset icdar2017. And it also has good performance in GPU occupancy, reasoning speed, and other indicators. The text detection results of actual transformer pictures show that the proposed algorithm solves the problem of poor detection performance of existing deep learning networks in dense text and long text of transformer pictures. |
| first_indexed | 2024-04-24T09:11:39Z |
| format | Article |
| id | doaj.art-817a7ac8ec7741f7bcbdc569ae82ff0d |
| institution | Directory Open Access Journal |
| issn | 1330-3651 1848-6339 |
| language | English |
| last_indexed | 2024-04-24T09:11:39Z |
| publishDate | 2022-01-01 |
| publisher | Faculty of Mechanical Engineering in Slavonski Brod, Faculty of Electrical Engineering in Osijek, Faculty of Civil Engineering in Osijek |
| record_format | Article |
| series | Tehnički Vjesnik |
| spelling | doaj.art-817a7ac8ec7741f7bcbdc569ae82ff0d2024-04-15T17:39:07ZengFaculty of Mechanical Engineering in Slavonski Brod, Faculty of Electrical Engineering in Osijek, Faculty of Civil Engineering in OsijekTehnički Vjesnik1330-36511848-63392022-01-0129386186610.17559/TV-20211027110610Text Detection of Transformer Based on Deep Learning AlgorithmYu Cheng0Yiru Wan1Yingjie Sima2Yinmei Zhang3Sanying Hu4Shu Wu5Hangzhou power supply company of State Grid Zhejiang Electric Power Co. Ltd., Marketing technology center, Zhejiang Province, Hangzhou, ChinaHangzhou power supply company of State Grid Zhejiang Electric Power Co. Ltd., Marketing technology center, Zhejiang Province, Hangzhou, ChinaState grid Zhejiang jiande power supply Co. Ltd, No. 288 Xin'an Road, Xinanjiang Street, Jiande city, Zhejiang Province, ChinaHangzhou power supply company of State Grid Zhejiang Electric Power Co. Ltd., Marketing technology center, Zhejiang Province, Hangzhou, ChinaHangzhou power supply company of State Grid Zhejiang Electric Power Co. Ltd., Marketing technology center, Zhejiang Province, Hangzhou, ChinaBeijing University of Posts and Telecommunications, 10 Xitucheng Road, Haidian District, Beijing, ChinaTransformers are important equipment in the power system. At present, the text information collection of transformer nameplates is through manual, which is inefficient. Therefore, it is necessary to find a high-precision automatic detection method of transformer text information. However, the current text detection algorithms have limited ability to detect special characters on the transformer. And they will also have the problem of incomplete detection in detecting the dense text and long text on the transformer nameplate. We propose a text detection network based on segmentation to automatically calibrate the text box of transformer nameplates. Our network is based on DB (differential binarization) network. It has a new feature fusion structure, which refers to the feature fusion structure of the u-net network. The proposed network has achieved better performance than the advanced scene text detection algorithms (DB, East) on the English scene text dataset icdar2015 and the Chinese-English mixed scene text dataset icdar2017. And it also has good performance in GPU occupancy, reasoning speed, and other indicators. The text detection results of actual transformer pictures show that the proposed algorithm solves the problem of poor detection performance of existing deep learning networks in dense text and long text of transformer pictures.https://hrcak.srce.hr/file/398877deep learningfeature fusiontext detection network based on classificationtransformer text detection |
| spellingShingle | Yu Cheng Yiru Wan Yingjie Sima Yinmei Zhang Sanying Hu Shu Wu Text Detection of Transformer Based on Deep Learning Algorithm Tehnički Vjesnik deep learning feature fusion text detection network based on classification transformer text detection |
| title | Text Detection of Transformer Based on Deep Learning Algorithm |
| title_full | Text Detection of Transformer Based on Deep Learning Algorithm |
| title_fullStr | Text Detection of Transformer Based on Deep Learning Algorithm |
| title_full_unstemmed | Text Detection of Transformer Based on Deep Learning Algorithm |
| title_short | Text Detection of Transformer Based on Deep Learning Algorithm |
| title_sort | text detection of transformer based on deep learning algorithm |
| topic | deep learning feature fusion text detection network based on classification transformer text detection |
| url | https://hrcak.srce.hr/file/398877 |
| work_keys_str_mv | AT yucheng textdetectionoftransformerbasedondeeplearningalgorithm AT yiruwan textdetectionoftransformerbasedondeeplearningalgorithm AT yingjiesima textdetectionoftransformerbasedondeeplearningalgorithm AT yinmeizhang textdetectionoftransformerbasedondeeplearningalgorithm AT sanyinghu textdetectionoftransformerbasedondeeplearningalgorithm AT shuwu textdetectionoftransformerbasedondeeplearningalgorithm |