摘要
绝缘子自爆缺陷检测是无人机巡检的重要内容。针对绝缘子缺陷区域场景复杂,形状尺寸不一,正负样本不均衡的问题,提出一种基于改进SSD的绝缘子自爆缺陷检测网络模型,用于绝缘子的缺陷检测。该模型引入一种CBAM空间通道注意力模块,提升了特征提取网络的学习能力,然后通过重构FPN特征金字塔结构,使用特征融合的方法提取多尺度特征缺陷,提高网络的特征提取能力。该模型还使用Focal Loss损失函数,用来解决SSD模型正负样本不均衡问题。经过实验验证,相对于其他模型,提出的网络改进模型F1值更高,针对绝缘子缺陷检测识别效果良好,检测速度能够满足实用要求。
Insulator self⁃detonation defect detection is an important part of UAV inspection.In view of the complex scenes,different shapes and sizes,and unbalanced positive and negative samples in insulator defect areas,an insulator self⁃detonation defect detection network model based on improved SSD(single shot MultiBox detector)is designed for insulator defect detection.In this model,a CBAM(convolutional block attention module)spatial channel attention module is introduced to enhance the learning ability of feature extraction network.And then,by reconstructing FPN(feature pyramid network)feature pyramid structure,multi⁃scale feature defects are extracted with the feature fusion method to improve the feature extraction ability of the network.In addition,Focal Loss is used to solve the problem of unbalanced positive and negative samples of SSD model.After test verification,in comparison with other models,the improved network model proposed in this paper has a higher F1 value,good detection and identification effect for insulator defects,and its detection speed can meet practical requirements.
作者
方毅
蒋作
FANG Yi;JIANG Zuo(School of Electrical and Information Technology,Yunnan Minzu University,Kunming 650504,China;School of Mathematics and Computer Science,Yunnan Minzu University,Kunming 650504,China)
出处
《现代电子技术》
2023年第15期49-54,共6页
Modern Electronics Technique
基金
国家自然科学基金项目(61866040)。
