无人机巡检图像和蒲公英算法的输电线缆绝缘子缺陷自动检测  

Unmanned aerial vehicle inspection image and dandelion algorithm for automatic transmission cable insulator defect detection

作　　者：黄小龙 HUANG Xiaolong(Sate Grid Jibei Electric Power Company Limited,Beijing 100052,China)

机构地区：[1]国网冀北电力有限公司,北京100052

出　　处：《集成电路与嵌入式系统》2024年第8期64-71,共8页Integrated Circuits and Embedded Systems

摘　　要：为提升输电线缆绝缘子缺失识别的准确度和速度,设计了基于无人机巡检图像和蒲公英算法(Dandelion Algorithm,DA)的输电线缆绝缘子缺陷自动检测系统。利用无人机检测技术实现对输电线缺失绝缘子的定位和检测,对无人机获取的绝缘子自动识别图像进行特征提取,规划无人机巡检路径,计算输电线缺失绝缘子的特征向量,利用最小二乘法拟合处理输电线缺失绝缘子的特征图像,通过计算绝缘子的位置门限确定漏损绝缘子的具体位置,从而可以精确确定缺陷的位置。在该过程中,引用了一种蒲公英算法用于无人机PID控制参数的优化问题,在此基础上,提出了一种基于遗传算法的种子播种方法,并对其进行了理论分析。实验结果表明,本文所提出的研究方法对物体辨识的正确率可达98%,具有较高的实用价值。In order to improve the accuracy and speed of missing insulator identification,an automatic insulator defect detection system based on UAV inspection images and Dandelion Algorithm(DA)is designed.UAV detection technology is used to locate and detect missing insulators of power transmission lines.Then,feature extraction is carried out on the automatic identification image of insulators obtained by UAV,and the operation and inspection path of UAV are planned.The feature vector of missing insulators of power lines is calculated.The feature image of missing insulators of power lines is processed by fitting the least square method.Obtain the location of the missing insulator to achieve accurate location of the defect.In this process,a Dandelion Algorithm is introduced to optimize the PID control parameters of UAVs.The seeding behavior and mathematical model of dandelion are described in principle,and the parameters are optimized by this algorithm.The test results show that the accuracy rate of target recognition is up to 98%,indicating that the research system is highly practical.

关 键 词：GPS DA算法 最小二乘法 绝缘子识别 PID控制 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]