Federated knowledge distillation for enhanced insulator defect detection in resource‐constrained environments

Abstract Insulator defect detection is crucial for the stable operation of power systems. It has become a mainstream research direction to realise insulator defect detection based on the combination of line images captured by UAVs and deep learning techniques. However, the existing high‐quality insu...

Full description

Saved in:
Bibliographic Details
Main Authors: Xiaohu Huang, Minghui Jia, Xianghua Tai, Wei Wang, Qi Hu, Dongping Liu, Peiheng Guo, Shengxiang Tian, Dequan Yan, Haishan Han
Format: Article
Language:English
Published: Wiley 2024-12-01
Series:IET Computer Vision
Subjects:
Online Access:https://doi.org/10.1049/cvi2.12290
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850058365805264896
author Xiaohu Huang
Minghui Jia
Xianghua Tai
Wei Wang
Qi Hu
Dongping Liu
Peiheng Guo
Shengxiang Tian
Dequan Yan
Haishan Han
author_facet Xiaohu Huang
Minghui Jia
Xianghua Tai
Wei Wang
Qi Hu
Dongping Liu
Peiheng Guo
Shengxiang Tian
Dequan Yan
Haishan Han
author_sort Xiaohu Huang
collection DOAJ
description Abstract Insulator defect detection is crucial for the stable operation of power systems. It has become a mainstream research direction to realise insulator defect detection based on the combination of line images captured by UAVs and deep learning techniques. However, the existing high‐quality insulator defect detection models still face problems such as relying on massive‐labelled data and huge model parameters. Especially on resource‐constrained devices, it becomes a challenge to strike a balance between model lightweighting and performance. Although the knowledge distillation technique provides a solution for model lightweighting, the loss of information in the distillation process leads to the performance degradation of small models, which in turn creates a paradox between lightweighting and performance. Hence, an insulator defect detection method based on federated knowledge distillation is proposed. The method not only realises the lightweighting of the model, but also effectively improves the model performance by collaboratively training the model through the federated learning approach. Moreover, the asynchronous aggregation approach and model freshness mechanism designed in the method further enhance the training efficiency and collaborative effect. The experimental results show that the detection accuracy and efficiency of this paper's method on public datasets are significantly better than the benchmark algorithm.
format Article
id doaj-art-b82995fb3ece44ac809d6b71ed0e8c9d
institution DOAJ
issn 1751-9632
1751-9640
language English
publishDate 2024-12-01
publisher Wiley
record_format Article
series IET Computer Vision
spelling doaj-art-b82995fb3ece44ac809d6b71ed0e8c9d2025-08-20T02:51:11ZengWileyIET Computer Vision1751-96321751-96402024-12-011881072108610.1049/cvi2.12290Federated knowledge distillation for enhanced insulator defect detection in resource‐constrained environmentsXiaohu Huang0Minghui Jia1Xianghua Tai2Wei Wang3Qi Hu4Dongping Liu5Peiheng Guo6Shengxiang Tian7Dequan Yan8Haishan Han9State Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaState Grid Qinghai Electric Power Company Ultra High Voltage Company Xi Ning Qinghai ChinaAbstract Insulator defect detection is crucial for the stable operation of power systems. It has become a mainstream research direction to realise insulator defect detection based on the combination of line images captured by UAVs and deep learning techniques. However, the existing high‐quality insulator defect detection models still face problems such as relying on massive‐labelled data and huge model parameters. Especially on resource‐constrained devices, it becomes a challenge to strike a balance between model lightweighting and performance. Although the knowledge distillation technique provides a solution for model lightweighting, the loss of information in the distillation process leads to the performance degradation of small models, which in turn creates a paradox between lightweighting and performance. Hence, an insulator defect detection method based on federated knowledge distillation is proposed. The method not only realises the lightweighting of the model, but also effectively improves the model performance by collaboratively training the model through the federated learning approach. Moreover, the asynchronous aggregation approach and model freshness mechanism designed in the method further enhance the training efficiency and collaborative effect. The experimental results show that the detection accuracy and efficiency of this paper's method on public datasets are significantly better than the benchmark algorithm.https://doi.org/10.1049/cvi2.12290big datalearning (artificial intelligence)mobile computingobject detection
spellingShingle Xiaohu Huang
Minghui Jia
Xianghua Tai
Wei Wang
Qi Hu
Dongping Liu
Peiheng Guo
Shengxiang Tian
Dequan Yan
Haishan Han
Federated knowledge distillation for enhanced insulator defect detection in resource‐constrained environments
IET Computer Vision
big data
learning (artificial intelligence)
mobile computing
object detection
title Federated knowledge distillation for enhanced insulator defect detection in resource‐constrained environments
title_full Federated knowledge distillation for enhanced insulator defect detection in resource‐constrained environments
title_fullStr Federated knowledge distillation for enhanced insulator defect detection in resource‐constrained environments
title_full_unstemmed Federated knowledge distillation for enhanced insulator defect detection in resource‐constrained environments
title_short Federated knowledge distillation for enhanced insulator defect detection in resource‐constrained environments
title_sort federated knowledge distillation for enhanced insulator defect detection in resource constrained environments
topic big data
learning (artificial intelligence)
mobile computing
object detection
url https://doi.org/10.1049/cvi2.12290
work_keys_str_mv AT xiaohuhuang federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT minghuijia federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT xianghuatai federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT weiwang federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT qihu federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT dongpingliu federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT peihengguo federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT shengxiangtian federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT dequanyan federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments
AT haishanhan federatedknowledgedistillationforenhancedinsulatordefectdetectioninresourceconstrainedenvironments