Multi-Geometric Reasoning Network for Insulator Defect Detection of Electric Transmission Lines

Abstract

To address the challenges in the unmanned system-based intelligent inspection of electric transmission line insulators, this paper proposed a multi-geometric reasoning network (MGRN) to accurately detect insulator geometric defects based on aerial images with complex backgrounds and different scales. The spatial geometric reasoning sub-module (SGR) was developed to represent the spatial location relationship of defects. The appearance geometric reasoning sub-module (AGR) and the parallel feature transformation (PFT) sub-module were adopted to obtain the appearance geometric features from the real samples. These multi-geometric features can be fused with the original visual features to identify and locate the insulator defects. The proposed solution is assessed through experiments against the existing solutions and the numerical results indicate that it can significantly improve the detection accuracy of multiple insulator defects using the aerial images.

Context-Enhanced Network with Spatial-Aware Graph for Smartphone Screen Defect Detection

“…They applied this approach in domains such as steel and textiles. Zhai et al [ 26 ] focus on insulator defect detection, constructing both a multi-geometry reasoning network and an appearance-geometry reasoning network to extract appearance and geometry features of defect samples, respectively. However, these works ignore the spatial relationships between defect regions and implicit information in images.…”

Deep Power Vision Technology and Intelligent Vision Sensors

“…This Special Issue aims to provide some up-to-date solutions to the problems of inspection in the power system and offer helpful reference for further research of deep power vision technology and intelligent vision sensors. It includes ten papers covering the tasks of detection for the inspection of transmission lines and substation [ 1 , 2 , 3 , 4 , 5 , 6 , 7 ], classification related to the requirements of inspection [ 8 , 9 ], and image defogging for transmission lines [ 10 ].…”

“…Han et al [ 4 ] were concerned with the shortcomings of IoU as well as the sensitivity of small targets to the model regression accuracy and proposed an improved YOLOX to solve the problem of low accuracy of insulator defect detection. Zhai et al [ 5 ] introduced a multi-geometric reasoning network to accurately detect insulator geometric defects based on aerial images with complex backgrounds and different scales which significantly improved the detection accuracy of multiple insulator defects using aerial images. Xin et al [ 6 ] combined the defogging algorithm with a two-stage detection model in order to accomplish the accurate detection of the insulator umbrella disc shedding in foggy weather.…”

A Few-Shot Defect Detection Method for Transmission Lines Based on Meta-Attention and Feature Reconstruction

“…Subsequently, they fine-tuned the model using a small number of vibration damper and wire clamp samples to achieve few-shot detection. On the other hand, Zhai et al [23] pre-trained Faster R-CNN on artificial samples and fine-tuned it using insulator defects. Ultimately, they achieved a detection accuracy of 62.7% mAP for insulator defects with a real sample size of 184.…”