An Improved Algorithm for Insulator and Defect Detection Based on YOLOv4
To further improve the accuracy and speed of UAV inspection of transmission line insulator defects, this paper proposes an insulator detection and defect identification algorithm based on YOLOv4, which is called DSMH-YOLOv4. In the feature extraction network of the YOLOv4 model, the improved algorit...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-02-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/12/4/933 |
| _version_ | 1797621319933624320 |
|---|---|
| author | Gujing Han Qiwei Yuan Feng Zhao Ruijie Wang Liu Zhao Saidian Li Min He Shiqi Yang Liang Qin |
| author_facet | Gujing Han Qiwei Yuan Feng Zhao Ruijie Wang Liu Zhao Saidian Li Min He Shiqi Yang Liang Qin |
| author_sort | Gujing Han |
| collection | DOAJ |
| description | To further improve the accuracy and speed of UAV inspection of transmission line insulator defects, this paper proposes an insulator detection and defect identification algorithm based on YOLOv4, which is called DSMH-YOLOv4. In the feature extraction network of the YOLOv4 model, the improved algorithm improves the residual edges of the residual structure based on feature reuse and designs the backbone network D-CSPDarknet53, which greatly reduces the number of parameters and computation of the model. The SA-Net (Shuffle Attention Neural Networks) attention model is embedded in the feature fusion network to strengthen the attention of target features and improve the weight of the target. Multi-head output is added to the output layer to improve the ability of the model to recognize the small target of insulator damage. The experimental results show that the number of parameters of the improved algorithm model is only 25.98% of that of the original model, and the mAP (mean Average Precision) of the insulator and defect is increased from 92.44% to 96.14%, which provides an effective way for the implementation of edge end algorithm deployment. |
| first_indexed | 2024-03-11T08:55:14Z |
| format | Article |
| id | doaj.art-b90929e7380e41f5b6f5a5f50df6c120 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-11T08:55:14Z |
| publishDate | 2023-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-b90929e7380e41f5b6f5a5f50df6c1202023-11-16T20:12:18ZengMDPI AGElectronics2079-92922023-02-0112493310.3390/electronics12040933An Improved Algorithm for Insulator and Defect Detection Based on YOLOv4Gujing Han0Qiwei Yuan1Feng Zhao2Ruijie Wang3Liu Zhao4Saidian Li5Min He6Shiqi Yang7Liang Qin8School of Electronic and Electrical Engineering, Wuhan Textile University, Wuhan 430200, ChinaSchool of Electronic and Electrical Engineering, Wuhan Textile University, Wuhan 430200, ChinaState Grid Information & Telecommunication Group Co., Ltd., Beijing 102211, ChinaSchool of Electronic and Electrical Engineering, Wuhan Textile University, Wuhan 430200, ChinaSchool of Electronic and Electrical Engineering, Wuhan Textile University, Wuhan 430200, ChinaSchool of Electronic and Electrical Engineering, Wuhan Textile University, Wuhan 430200, ChinaSchool of Electrical and Automation, Wuhan University, Wuhan 430072, ChinaSchool of Electrical and Automation, Wuhan University, Wuhan 430072, ChinaSchool of Electrical and Automation, Wuhan University, Wuhan 430072, ChinaTo further improve the accuracy and speed of UAV inspection of transmission line insulator defects, this paper proposes an insulator detection and defect identification algorithm based on YOLOv4, which is called DSMH-YOLOv4. In the feature extraction network of the YOLOv4 model, the improved algorithm improves the residual edges of the residual structure based on feature reuse and designs the backbone network D-CSPDarknet53, which greatly reduces the number of parameters and computation of the model. The SA-Net (Shuffle Attention Neural Networks) attention model is embedded in the feature fusion network to strengthen the attention of target features and improve the weight of the target. Multi-head output is added to the output layer to improve the ability of the model to recognize the small target of insulator damage. The experimental results show that the number of parameters of the improved algorithm model is only 25.98% of that of the original model, and the mAP (mean Average Precision) of the insulator and defect is increased from 92.44% to 96.14%, which provides an effective way for the implementation of edge end algorithm deployment.https://www.mdpi.com/2079-9292/12/4/933UAV inspectioninsulator defectDSMH-YOLOv4feature reuseSA-Netmulti-head |
| spellingShingle | Gujing Han Qiwei Yuan Feng Zhao Ruijie Wang Liu Zhao Saidian Li Min He Shiqi Yang Liang Qin An Improved Algorithm for Insulator and Defect Detection Based on YOLOv4 Electronics UAV inspection insulator defect DSMH-YOLOv4 feature reuse SA-Net multi-head |
| title | An Improved Algorithm for Insulator and Defect Detection Based on YOLOv4 |
| title_full | An Improved Algorithm for Insulator and Defect Detection Based on YOLOv4 |
| title_fullStr | An Improved Algorithm for Insulator and Defect Detection Based on YOLOv4 |
| title_full_unstemmed | An Improved Algorithm for Insulator and Defect Detection Based on YOLOv4 |
| title_short | An Improved Algorithm for Insulator and Defect Detection Based on YOLOv4 |
| title_sort | improved algorithm for insulator and defect detection based on yolov4 |
| topic | UAV inspection insulator defect DSMH-YOLOv4 feature reuse SA-Net multi-head |
| url | https://www.mdpi.com/2079-9292/12/4/933 |
| work_keys_str_mv | AT gujinghan animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT qiweiyuan animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT fengzhao animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT ruijiewang animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT liuzhao animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT saidianli animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT minhe animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT shiqiyang animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT liangqin animprovedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT gujinghan improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT qiweiyuan improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT fengzhao improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT ruijiewang improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT liuzhao improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT saidianli improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT minhe improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT shiqiyang improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 AT liangqin improvedalgorithmforinsulatoranddefectdetectionbasedonyolov4 |