摘要
针对传统SF6气体泄漏检测方法存在图像采集和泄漏识别精度低的问题,提出设计一种基于DCNN网络的SF6气体在线监测系统。首先,采用OV78和MSP430单片机进行泄漏气体图像采集;然后通过WIFI无线通信将采集数据输入至深度学习模块中进行GMM泄漏区域提取和多特征提取;最后采用DCNN神经网络进行SF6气体泄漏准确识别和分类。实验结果表明,相较于传统的Lenet-5、ZF-net和Alexnet经典网络,提出的DCNN方法无论在网络性能,还是在识别的准确率方面,均具备良好的表现,其识别准确率最高可达82%,识别性能均优于另外三种网络模型。实际应用表明,该方法具有良好的检测效果。由此说明本文构建的系统可用于电力中的SF6气体在线监测,保障电力的安全。
Aiming at the problem of low image acquisition and leakage recognition accuracy in the traditional SF6 gas leakage detection method,a SF6 gas online monitoring system based on DCNN network is proposed.First,OV 78 and MSP430 are used for leakage gas image acquisition;then the collected data is input into the deep learning module through WIFI wireless communication for GMM leakage area extraction and multi-feature extraction;finally,DCNN neural network for accurate identification and classification of SF6 gas leakage.The experimental results show that compared with the traditional Lenet-5,ZF-net and Alexnet classical networks,the DCNN method has good performance and recognition accuracy of up to 82%,and the recognition performance is better than the other three network models.Practical application shows that the proposed method has a good detection effect.This shows that the system constructed in this paper can be used for SF6 gas online monitoring in electric power to ensure the safety of electric power.
作者
辛拓
谢欢欢
张宏钊
陈龙
何维
黄炜昭
XIN Tuo;XIE Huanhuan;ZHANG Hongzhao;CHEN Long;HE Wei;HUANG Weizhao(Shenzhen Power Supply Bureau Co.,Ltd.,Shenzhen,Guangdong,518000,China)
出处
《自动化与仪器仪表》
2023年第7期273-277,282,共6页
Automation & Instrumentation
基金
深圳供电局有限公司《支撑智能运检的新型变电站及装备模块化技术研究》(090000KK52200150)。
