电气传动2026，Vol.56Issue(5)：45-51,7.DOI:10.19457/j.1001-2095.dqcd26587
基于多源数据融合的电力作业机载感知设备设计
Design of an On-board Perception Device for Power Operations Based on Multi-source Data Fusion
摘要
Abstract
As the requirements for real-time performance and accuracy in power operation scenarios continue to rise,traditional single-source perception methods face significant challenges in complex environments and decision making support.To address this issue,the design of an on-board perception device for power operations based on multi-source data fusion technology was presented.The instrument integrates multi-source data from lidar,Realsense depth cameras,and stereo cameras,and combines multi-source data adaptive rotational convolution(MSDARC)and maximum probability distance intersection over union(MPDIoU)algorithms to significantly enhance target perception and real-time response.The system was applied in a power grid live-line operation robot,where it excelled in tasks such as point cloud stitching and wire recognition,with the maximum time for point cloud stitching not exceeding 0.14 s and an average precision(AP)of 0.852.These results demonstrate the potential of the system for use in complex environments.This study provides theoretical support for intelligent perception in power operations and offers new ideas and methods for applying multi-source data fusion technology in other fields.关键词
多源数据融合/电力作业/机载感知设备/多源数据自适应旋转卷积/引线识别Key words
multi-source data fusion/power operation/on-board perception device/multi-source data adaptive rotational convolution(MSDARC)/wire recognition分类
信息技术与安全科学引用本文复制引用
胡益菲,魏然,陈志华,李威,史成亮..基于多源数据融合的电力作业机载感知设备设计[J].电气传动,2026,56(5):45-51,7.基金项目
国网天津市电力公司科技项目(滨海-研发2024-01) （滨海-研发2024-01）
