[1]陈炳煌,缪希仁,江灏,等.融合粒子群与极限学习机的输电杆塔灾害分类方法[J].郑州大学学报(工学版),2021,42(04):77-83.[doi:10.13705/j.issn.1671-6833.2021.04.004]
　Chen Binghuang,Miao Xiren,Jiang Yan,et al.A Method for Disaster Status Classification of Transmission Line Towers by Integrating Particle Swarm Optimization and Extreme Learning Machine[J].Journal of Zhengzhou University (Engineering Science),2021,42(04):77-83.[doi:10.13705/j.issn.1671-6833.2021.04.004]

点击复制

融合粒子群与极限学习机的输电杆塔灾害分类方法()

分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:

42

期数:

2021年04期

页码:

77-83

栏目:

出版日期:

2021-07-30

文章信息/Info

Title:

A Method for Disaster Status Classification of Transmission Line Towers by Integrating Particle Swarm Optimization and Extreme Learning Machine

作者:

陈炳煌; 缪希仁; 江灏; 吴俊钢

福州大学电气工程与自动化学院;福建工程学院电子电气与物理学院;

Author(s):

Chen Binghuang; Miao Xiren; Jiang Yan; Wu Jungang;

School of Electrical Engineering and Automation of Fuzhou University; School of Electronics and Physics, School of Electronics and Physics of Fujian Institute of Engineering;

关键词:

粒子群优化; 极限学习机; 输电线路杆塔; 灰色关联分析

Keywords:

particle swarm optimization; extreme learning machine; transmission line tower; gray correlation analysis

DOI:

10.13705/j.issn.1671-6833.2021.04.004

文献标志码:

A

摘要:

在无人机应急巡检架空输电线路的基础上，提出一种融合粒子群优化与极限学习机的输电线路铁塔灾害状态分类方法。先结合直线检测法和Harris角点检测法从架空输电线路铁塔图像中提取特征参数，再采用灰色关联分析法获取铁塔与灾害状态关联的主要特征参数，应用粒子群优化对极限学习机的输入隐藏权值和隐藏偏差阈值进行优化，将该分类方法应用于无人机铁塔图像数据集中。与其他算法的对比实验表明，融合粒子群优化和极限学习机模型的分类方法可更准确地区分正常、半倒塌和全倒塌等三种类型的输电线路铁塔状态。

Abstract:

The damage of natural disasters to transmission lines could affect the safety of power grid operation seriously.However,it would be difficult to evaluate the classification of transmission line towers accurately during emergency inspection by drone.Based on the emergency inspection of transmission lines by drone,this paper proposed a classification method of transmission line tower that integrated particle swarm optimization and extreme learning machine.Transmission line tower disaster state could be divided into three types:normal,half collapse and full collapse.Firstly,the disaster state image data set of the towers of the transmission line emergency inspection was established,and the tower contour and its 7 main characteristic parameters were extracted from the image data set by combining the linear segment detection and Harris corner detection method.Then,the grey relation analysis method was used to obtain the 4 key characteristic parameters associated with the tower image and the disaster state.Then,the classification accuracy of towers was taken as the fitness of particle swarm optimization algorithm,and the hidden weights and hidden deviation threshold were optimized by using particle swarm optimization algorithm.The weights were imported into the extreme learning machine to train the four key characteristic parameters of the tower disaster state images.Finally,it was applied to the emergency inspection to classify the disaster state images of transmission line towers.It was found that the four key characteristic parameters of transmission line tower image and disaster status classification were circularity,length-width ratio,rectangularity and relative position of gravity center.The experimental results showed that the classification accuracy of the fusion particle swarm optimization and extreme learning machine model was 88.33%,and the accuracy rate was 92.68%.Compared with the backpropagation neural network and support vector machine algorithm model,it had better detection and classification effect.At the same time,the feasibility and effectiveness of the classification method based on particle swarm optimization and extreme learning machine were verified.

常用功能