摘要
图像分类算法常被搭载在无人机系统中,以剔除无人机巡线过程中采集到的大量无用数据。针对这一问题,本文在分析了无用图像及有用图像特征的基础上,提出了一种基于卷积神经网络的输电线路智能图像筛选方法。收集无人机巡检捕捉到的航拍图像,并以此为基础建立了一个输电线路航拍数据集,基于ResNet优化并利用航拍图像数据集训练该网络,经过多次迭代训练保留最优权重,通过加载最优权重的神经网络对测试样本进行分类,精确度达到99.00%。
Image classification algorithm is often carried in UAV system to eliminate a large number of useless data collected in the process of UAV based line patrol.To solve this problem,this paper proposes an intelligent image filtering method based on the convolutional neural network.Firstly,the aerial images captured by cameras are collected during UAV based line patrol,establishing an aerial image dataset of transmission lines.Then,an optimized network based on the ResNet is proposed and trained by aerial image dataset,and the best weight is loaded and utilized to classify the test specimens.The accuracy has reached 99.00%.
作者
李弘宸
杨忠
姜遇红
韩家明
赖尚祥
张秋雁
LI Hongchen;YANG Zhong;JIANG Yuhong;HAN Jiaming;LAI Shangxiang;ZHANG Qiuyan(College of Automation Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China;Research Institute of UAV,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China;Electric Power Research Institute of Guizhou Power Grid Co.,Ltd.,Guiyang 550000,China)
出处
《应用科技》
CAS
2021年第2期64-68,共5页
Applied Science and Technology
基金
国家自然科学基金项目(61473144)
贵州省科技计划项目(黔科合支撑[2020]2Y044号)
中国南方电网有限责任公司科技项目(066600KK52170074)
江苏高校优势学科建设工程项目
南京航空航天大学研究生创新基地(实验室)开放基金项目(kfjj20190305).
关键词
无人机巡线
输电线路
深度学习
图像分类
卷积神经网络
ResNet
分组卷积
网络轻量化
UAV based line patrol
transmission lines
deep learning
image classification
convolutional neural network
ResNet
group convolution
network lightweight
