Xiaoming Lei, Mengjin Sun, Rongxin Zhao, Huayong Wu, Zijie Zhou, You Dong, Limin Sun
{"title":"Unsupervised Vision-Based Structural Anomaly Detection and Localization with Reverse Knowledge Distillation","authors":"Xiaoming Lei, Mengjin Sun, Rongxin Zhao, Huayong Wu, Zijie Zhou, You Dong, Limin Sun","doi":"10.1155/2024/8933148","DOIUrl":null,"url":null,"abstract":"<div>\n <p>Most of vision-based methods for structural damage detection rely on supervised learning, requiring a substantial number of labeled images for model training, which is labor-intensive and time-consuming. To address these challenges, this study introduces a vision-based structural anomaly detection and localization approach using unsupervised learning and reverse knowledge distillation. The proposed model incorporates a teacher model, a student model, and a trainable one-class bottleneck embedding module. The asymmetrical architecture of the teacher and student models forms an encoder-decoder structure for parameter transfer and feature extraction. The student network receives a specific embedding from the teacher network as input and target, facilitating the recovery of multiscale information from the teacher. Training images only contain the undamaged structures, and the teacher model, a pretrained model, instructs the student model to remember their undamaged features to detect and localize damages in unseen testing images. Through experiments, including a comparison among five candidate backbones for pretrained teacher models based on the residual network and testing across various structural damage types, the optimal model is identified, demonstrating good performance in both anomaly detection and localization. Furthermore, the model’s generalization performance is thoroughly validated, confirming its efficacy across diverse scenarios.</p>\n </div>","PeriodicalId":49471,"journal":{"name":"Structural Control & Health Monitoring","volume":null,"pages":null},"PeriodicalIF":4.6000,"publicationDate":"2024-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1155/2024/8933148","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Structural Control & Health Monitoring","FirstCategoryId":"5","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1155/2024/8933148","RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CONSTRUCTION & BUILDING TECHNOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
Most of vision-based methods for structural damage detection rely on supervised learning, requiring a substantial number of labeled images for model training, which is labor-intensive and time-consuming. To address these challenges, this study introduces a vision-based structural anomaly detection and localization approach using unsupervised learning and reverse knowledge distillation. The proposed model incorporates a teacher model, a student model, and a trainable one-class bottleneck embedding module. The asymmetrical architecture of the teacher and student models forms an encoder-decoder structure for parameter transfer and feature extraction. The student network receives a specific embedding from the teacher network as input and target, facilitating the recovery of multiscale information from the teacher. Training images only contain the undamaged structures, and the teacher model, a pretrained model, instructs the student model to remember their undamaged features to detect and localize damages in unseen testing images. Through experiments, including a comparison among five candidate backbones for pretrained teacher models based on the residual network and testing across various structural damage types, the optimal model is identified, demonstrating good performance in both anomaly detection and localization. Furthermore, the model’s generalization performance is thoroughly validated, confirming its efficacy across diverse scenarios.
期刊介绍:
The Journal Structural Control and Health Monitoring encompasses all theoretical and technological aspects of structural control, structural health monitoring theory and smart materials and structures. The journal focuses on aerospace, civil, infrastructure and mechanical engineering applications.
Original contributions based on analytical, computational and experimental methods are solicited in three main areas: monitoring, control, and smart materials and structures, covering subjects such as system identification, health monitoring, health diagnostics, multi-functional materials, signal processing, sensor technology, passive, active and semi active control schemes and implementations, shape memory alloys, piezoelectrics and mechatronics.
Also of interest are actuator design, dynamic systems, dynamic stability, artificial intelligence tools, data acquisition, wireless communications, measurements, MEMS/NEMS sensors for local damage detection, optical fibre sensors for health monitoring, remote control of monitoring systems, sensor-logger combinations for mobile applications, corrosion sensors, scour indicators and experimental techniques.