摘要详情

ID / 提交时间

5 / 2023-08-18 15:05:11

标题

Insulator Defect Detection based on Faster R-CNN and YOLOv3 Algorithm

关键字

Insulator; Faster R-CNN; YOLOv3; EfficientNet-B3; CBAM

主题及专题

6.AI-driven technology > (1)、Intelligent Anomaly Detection, Fault Diagnosis and Prognostics

状态

终稿

作者

Ping Hu / School of Big Health and Intelligent Engineering, Chengdu Medical College

Junchen Lu / School of Big Health and Intelligent Engineering, Chengdu Medical College

Yuan Cui / School of Big Health and Intelligent Engineering, Chengdu Medical College

Bo Hu / School of Big Health and Intelligent Engineering, Chengdu Medical College

Fan Liang / Tangshan Research Institute,Southwest Jiaotong University

摘要

In order to solve the problem that the existing insulator defect detection models are not effective enough for the tasks such as complex background and low signal-to-noise ratio of the target objects, this paper proposes an insulator defect detection method based on Faster R-CNN and YOLOv3 target detection algorithm. First, operations such as random angle rotation, flipping, adjusting contrast, and adding noise are used to preprocess the data images. Secondly, the Faster R-CNN algorithm is used as the basis to realize the accurate localization of insulators. Finally, EfficientNet-B3 is used as the backbone network of YOLOv3, and the dual-attention mechanism CBAM is embedded to realize the insulator defect recognition. The results show that compared with the existing models, the insulator defect detection method proposed in this paper exhibits more accurate insulator localization and defect recognition performance.