A Lightweight Insulator Detection Methodology for UAVs in Power Line Inspection
-
Published:2023-09-02
Issue:
Volume:
Page:
-
ISSN:0218-1266
-
Container-title:Journal of Circuits, Systems and Computers
-
language:en
-
Short-container-title:J CIRCUIT SYST COMP
Author:
Meng Yue1ORCID,
Tang Yiming1ORCID,
Huang Xiang1ORCID,
Wang Haoyu1ORCID,
Zhu Jie1ORCID,
Tang Wenjuan1ORCID,
Chen Lu1ORCID
Affiliation:
1. Jiangsu Frontier Electric Power Technology Co., Ltd., Nanjing, Jiangsu, P. R. China
Abstract
The defect of insulators on the power line may lead to the failure of the electric transmission system, which in turn endangers the safe and reliable operation of the whole power system. Therefore, timely and accurate detection of insulators has become a current research hotspot. However, manual inspection cannot guarantee real-time detection and is prone to security accidents. Thanks to the convenience of Unmanned Aerial Vehicles (UAVs), collecting insulator images by UAVs and designing detection algorithms on insulators for real-time and accurate insulator detection have become a practical option. In this paper, considering the poor computing resources of UAVs, a lightweight object detection algorithm, NanoDet, is trained to detect insulators. In addition, to fully utilize the computing resources, the proposed model is segmented. Power-efficient layers are mapped to the central processing units (CPUs) for execution. Experimental results indicate that the proposed model can cut down the power consumption by up to 46.4% without violating the time constraint.
Funder
Science and Technology Project of Jiangsu Frontier Electric Power Technology Co., Ltd.
Publisher
World Scientific Pub Co Pte Ltd
Subject
Electrical and Electronic Engineering,Hardware and Architecture,Electrical and Electronic Engineering,Hardware and Architecture