ID-Det: Insulator Burst Defect Detection from UAV Inspection Imagery of Power Transmission Facilities DOI Creative Commons
Shangzhe Sun, Chi Chen, Bisheng Yang

et al.

Drones, Journal Year: 2024, Volume and Issue: 8(7), P. 299 - 299

Published: July 5, 2024

The global rise in electricity demand necessitates extensive transmission infrastructure, where insulators play a critical role ensuring the safe operation of power systems. However, are susceptible to burst defects, which can compromise system safety. To address this issue, we propose an insulator defect detection framework, ID-Det, comprises two main components, i.e., Insulator Segmentation Network (ISNet) and Burst Detector (IBD). (1) ISNet incorporates novel Clipping Module (ICM), enhancing segmentation performance. (2) IBD leverages corner extraction methods periodic distribution characteristics corners, facilitating key corners on mask accurate localization defects. Additionally, construct Defect Dataset (ID Dataset) consisting 1614 images. Experiments dataset demonstrate that ID-Det achieves accuracy 97.38%, precision recall rate 94.56%, outperforming general with 4.33% increase accuracy, 5.26% precision, 2.364% recall. also shows 27.2% improvement Average Precision (AP) compared baseline. These results indicate has significant potential for practical application inspection.

Language: Английский

Development of an Aerial Manipulation System Using Onboard Cameras and a Multi-Fingered Robotic Hand with Proximity Sensors DOI Creative Commons
Ryuki Sato,

Etienne Marco Badard,

C. S. Thaymara Romulo

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(2), P. 470 - 470

Published: Jan. 15, 2025

Recently, aerial manipulations are becoming more and important for the practical applications of unmanned vehicles (UAV) to choose, transport, place objects in global space. In this paper, an manipulation system consisting a UAV, two onboard cameras, multi-fingered robotic hand with proximity sensors is developed. To achieve self-contained autonomous navigation targeted object, tracking depth cameras used detect object control UAV reach target even Global Positioning System-denied environment. The can perform sensor-based grasping stably that within position error tolerance (a circle radius 50 mm) from center hand. Therefore, successfully grasp requirement (=UAV) during hovering after reaching should be less than tolerance. meet requirement, detection algorithm support accurate localization by combining information both was addition, camera mount orientation attitude sampling rate were determined experiments, it confirmed these implementations improved robot Finally, experiments on using developed demonstrated successful object.

Language: Английский

Citations

0
ID-Det: Insulator Burst Defect Detection from UAV Inspection Imagery of Power Transmission Facilities DOI Creative Commons
Shangzhe Sun, Chi Chen, Bisheng Yang

et al.

Drones, Journal Year: 2024, Volume and Issue: 8(7), P. 299 - 299

Published: July 5, 2024

The global rise in electricity demand necessitates extensive transmission infrastructure, where insulators play a critical role ensuring the safe operation of power systems. However, are susceptible to burst defects, which can compromise system safety. To address this issue, we propose an insulator defect detection framework, ID-Det, comprises two main components, i.e., Insulator Segmentation Network (ISNet) and Burst Detector (IBD). (1) ISNet incorporates novel Clipping Module (ICM), enhancing segmentation performance. (2) IBD leverages corner extraction methods periodic distribution characteristics corners, facilitating key corners on mask accurate localization defects. Additionally, construct Defect Dataset (ID Dataset) consisting 1614 images. Experiments dataset demonstrate that ID-Det achieves accuracy 97.38%, precision recall rate 94.56%, outperforming general with 4.33% increase accuracy, 5.26% precision, 2.364% recall. also shows 27.2% improvement Average Precision (AP) compared baseline. These results indicate has significant potential for practical application inspection.

Language: Английский

Citations

3