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

Alahyari, Arman; Hinneck, Anton; Tariverdizadeh, Rahim; Pozo, David

doi:10.1109/sges51519.2020.00090

Cited by 9 publications

(7 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our proposed model achieved notable performance metrics with a recall of 0.92 and an F1 score of 0.94, indicating its high accuracy and reliability in identifying and classifying insulator faults. This performance surpasses that of other significant models in the literature, such as [25]- [28], Faster R-CNN, and various deep learning-based approaches that focus on spatial or temperature features (Table 5). Results in the study [28] achieved an F1 score of 1 but a lower recall of 0.8.…”

Section: Discussionmentioning

confidence: 76%

“…This technique has proven to be effective in identifying faults in both glass and ceramic insulators, demonstrating its robustness across various backgrounds. Another study [28] introduces a two-stage deep learning model that has been introduced for insulator fault detection. The initial stage of this model involves the segmentation of insulators from their backgrounds, employing a UNet architecture.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Infrared Thermography Based Insulator Fault Classification via Unsupervised Clustering and Semi-Supervised Learning

Shafique,

Alam,

Rashid

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Power substations play a crucial role in ensuring the reliable transmission of electricity to residential and commercial establishments. This paper addresses the critical issue of insulator fault detection in electric substations, emphasizing the importance of timely identification to prevent accidents. Conventional inspection methods relying on manual intervention pose high risk, especially in high-voltage substations. To overcome this, infra-red (IR) thermal cameras are employed, but human analysis of IR thermal images is subjective and time-consuming. The current paper develops a computer vision system uses advanced segmentation algorithms to automatically detect and categorize defects in insulators based on Infrared Thermography (IRT) images. The proposed system combines unsupervised clustering for efficient training with supervised learning for accurate classification. In particular, it first employs clustering techniques for unsupervised learning to reduce human labeling needs. Then, in the supervised learning phase, a Gaussian kernel support vector machine (SVM) algorithm classifies various insulator defects using extracted features. The trained algorithm utilizes the extracted characteristics to effectively identify and classify several fault types in thermal images of insulators. The system's efficiency in quickly identifying faults enables timely intervention, preventing potential power outages and safety hazards.

show abstract

Section: Discussionmentioning

confidence: 76%

Section: Related Workmentioning

confidence: 99%

Infrared Thermography Based Insulator Fault Classification via Unsupervised Clustering and Semi-Supervised Learning

Shafique,

Alam,

Rashid

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Alahyari et al. [65] used a two‐stage model for both the segmentation and the detection tasks of faulty insulators. The segmentation model achieved a total of 78% accuracy, while the classifier obtained 92%.…”

Section: Resultsmentioning

confidence: 99%

“…In this application only 1.38 MB of memory was required, making this a promising solution to be applied in an embedded system. Alahyari et al [65] used a twostage model for both the segmentation and the detection tasks of faulty insulators. The segmentation model achieved a total of 78% accuracy, while the classifier obtained 92%.…”

Section: Comparison To Related Studiesmentioning

confidence: 99%

Optimized hybrid YOLOu‐Quasi‐ProtoPNet for insulators classification

Stefenon

Singh

Souza

et al. 2023

IET Generation Trans & Dist

View full text Add to dashboard Cite

To ensure the electrical power supply, inspections are frequently performed in the power grid. Nowadays, several inspections are conducted considering the use of aerial images since the grids might be in places that are difficult to access. The classification of the insulators' conditions recorded in inspections through computer vision is challenging, as object identification methods can have low performance because they are typically pre‐trained for a generalized task. Here, a hybrid method called YOLOu‐Quasi‐ProtoPNet is proposed for the detection and classification of failed insulators. This model is trained from scratch, using a personalized ultra‐large version of YOLOv5 for insulator detection and the optimized Quasi‐ProtoPNet model for classification. For the optimization of the Quasi‐ProtoPNet structure, the backbones VGG‐16, VGG‐19, ResNet‐34, ResNet‐152, DenseNet‐121, and DenseNet‐161 are evaluated. The F1‐score of 0.95165 was achieved using the proposed approach (based on DenseNet‐161) which outperforms models of the same class such as the Semi‐ProtoPNet, Ps‐ProtoPNet, Gen‐ProtoPNet, NP‐ProtoPNet, and the standard ProtoPNet for the classification task.

show abstract

“…is method does not require an object detection network in advance and directly uses the semantic segmentation method to complete the identification and segmentation of insulators. Alahyari et al [15] proposed a twostage convolutional neural network model consisting of segmentation and classification units for fault classification of insulators. Aerial insulator images have the characteristics of complex backgrounds, many pseudotargets, and low signal-to-noise ratios, which bring difficulties for segment insulators.…”

Section: Introductionmentioning

confidence: 99%

Insulator Semantic Segmentation in Aerial Images Based on Multiscale Feature Fusion

2022

View full text Add to dashboard Cite

As one of the important components in the transmission line, the insulator is related to the safe and reliable operation of the entire transmission line. Aerial images are characterized by complex backgrounds, multiple pseudotargets, and low signal-to-noise ratios. Rapid and accurate localization of insulators in aerial images is a critical and challenging task in automatic inspection of transmission lines. Most insulator localization methods suffer from the loss of target edge detail information and large amount of model parameters. To solve these problems, this paper adopts an Encoder-Decoder architecture, called ED-Net, to realize end-to-end intelligent and accurate identification of insulators in aerial images. Firstly, Initial Module and CA-Bottleneck which are used to extract features from images to generate finer feature maps are proposed in the Encoder path. Meanwhile, global average pooling is used to preserve the maximum receptive field. Secondly, in the Decoder path, Refinement Boundary Module and Asymmetric Convolution Module are given to perform boundary optimization on the feature map, which are generated by the Encoder path. Finally, the Attention Feature Fusion Module is introduced into the Decoder path to combine high-level features with low-level features better and reduce the gap between features of different levels. The proposed model architecture keeps a suitable balance between the model parameters and insulator segmentation performance on insulator test datasets. Specifically, for a 512 × 512 input image, 95.12% mean intersection over union is achieved on the insulator test datasets with different environments and model parameters size being only 13.61 M. Compared with the current state-of-the-art semantic segmentation methods, the results show that the proposed method has higher efficient and accuracy.

show abstract

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

Cited by 9 publications

References 15 publications

Infrared Thermography Based Insulator Fault Classification via Unsupervised Clustering and Semi-Supervised Learning

Infrared Thermography Based Insulator Fault Classification via Unsupervised Clustering and Semi-Supervised Learning

Optimized hybrid YOLOu‐Quasi‐ProtoPNet for insulators classification

Insulator Semantic Segmentation in Aerial Images Based on Multiscale Feature Fusion

Contact Info

Product

Resources

About