Abstract:Inspection of insulators is important to ensure reliable operation of the power system. Deep learning is being increasingly exploited to automate the inspection process by leveraging object detection models to analyse aerial images captured by drones. A purely object detection-based approach, however, suffers from class imbalance-induced poor performance, which can be accentuated for infrequent and hard-to-detect incipient faults. This article proposes the use of anomaly detection along with object detection in a two-stage approach for incipient fault detection in a data-efficient manner. An explainable convolutional one-class classifier is adopted for anomaly detection. The one-class formulation reduces the reliance on plentifully available images of faulty insulators, while the explainability of the model is expected to promote adoption by the industry. A modified loss function is developed that addresses computational and interpretability issues with the existing model, also allowing for the integration of other losses. The superiority of the novel loss function is demonstrated with MVTec-AD dataset. The models are trained for insulator inspection with two datasets -- representing data-abundant and data-scarce scenarios -- in unsupervised and semi-supervised settings. The results suggest that including as few as five real anomalies in the training dataset significantly improves the model's performance and enables reliable detection of rarely occurring incipient faults in insulators.

What problem does this paper attempt to address?

### Problems Addressed by the Paper The paper aims to address several key issues in the detection of power line insulators: 1. **Class Imbalance Problem**: In practical applications, the number of faulty insulators is much smaller than that of normal insulators, leading to a class imbalance in the training dataset. This imbalance can result in poor performance of deep neural networks in detecting faulty insulators. 2. **Small-Scale Fault Patterns**: Fault patterns (such as flashover and breakage) usually occupy only a small part of the image, making the detection task more challenging. Existing object detection-based methods perform poorly in such cases. 3. **Insufficient Dataset Scale and Diversity**: Many existing studies use small-scale datasets that do not adequately reflect the complexity and diversity of real-world application scenarios, limiting the model's generalization ability. 4. **Simple Detection Tasks**: Existing research mainly focuses on detecting simpler tasks such as missing caps or plates, with less attention to early faults (such as flashover and breakage). These early faults are crucial for the maintenance and safe operation of the power grid. ### Solutions To address the above issues, the paper proposes a two-stage method that combines object detection and anomaly detection techniques for the automated detection of early faults in power line insulators: 1. **Two-Stage Method**: - **Stage 1**: Use an object detection model to locate insulators and plates. - **Stage 2**: Use an anomaly detection model to classify the extracted plate images, identifying healthy and faulty plates. 2. **Improved Anomaly Detection Model**: - Employ an interpretable Fully Convolutional Data Description (FCDD) for anomaly detection. - Introduce a new loss function that addresses computational and interpretability issues of existing models and allows integration with other loss functions. 3. **Data Efficiency**: - Significantly improve model performance by using a small number of real fault samples, enabling reliable detection of rare early faults. 4. **Experimental Validation**: - Conduct experiments using two datasets (one data-rich and one data-scarce) to validate the effectiveness of the proposed method. Through this approach, the paper not only addresses the class imbalance and small-scale fault pattern issues but also enhances model interpretability and data utilization efficiency, providing a new solution for the automated detection of power line insulators.

An Improved Anomaly Detection Model for Automated Inspection of Power Line Insulators

Object detection-based inspection of power line insulators: Incipient fault detection in the low data-regime

XAI-guided Insulator Anomaly Detection for Imbalanced Datasets

Anomaly Detection Algorithm Based on Electric Equipment

Automatic Insulator Detection for Power Line Using Aerial Images Powered by Convolutional Neural Networks

A Transfer Learning Architecture to Detect Faulty Insulators in Powerlines

Catenary insulator defect detection based on contour features and gray similarity matching

Detection of Power Line Insulator Defects Using Aerial Images Analyzed With Convolutional Neural Networks

A deep learning approach for insulator instance segmentation and defect detection

Anomaly detection for industrial quality assurance: A comparative evaluation of unsupervised deep learning models

Deep learning based insulator fault detection algorithm for power transmission lines

Insulator defect detection algorithm based on adaptive feature fusion and lightweight YOLOv5s

Anomaly detection in IR images of PV modules using supervised contrastive learning

Insulator Anomaly Detection Method Based on Few-Shot Learning

SODD-YOLOv8: An Insulator Defect Detection Algorithm based on Feature Enhancement and Variable Row Convolution

Research on the application of lightweight deep learning model in automatic identification of electricity-related public safety hazards in distribution lines

Pre-Trained Large Model Based Automatic Generation of Defect Insulator Images for Power Distribution Systems

Detection of Glass Insulators Using Deep Neural Networks Based on Optical Imaging

Anomaly Detection of High-Speed Railway Catenary Damage

Data Augmentation of Insulator Aerial Images for Intelligent Inspection of Electric Transmission Systems

Segmentation and Defect Classification of the Power Line Insulators: A Deep Learning-based Approach