基于卷积神经网络对磁异常信号的识别研究  

Research on Recognition of Magnetic Anomaly Signal Based on Convolutional Neural Network

作　　者：李启飞 吴芳 林义杰 LI Qifei;WU fang;LINYijie(Naval Aviation University,Yantai Shandong 264001,China;The 91550th Unit of PLA,Dalian Liaoning 116000,China;The 92485th Unit of PLA,Dalian Liaoning 116113,China)

机构地区：[1]海军航空大学,山东烟台264001 [2]91550部队,辽宁大连116000 [3]92485部队,辽宁大连116113

出　　处：《海军航空工程学院学报》2020年第2期161-166,共6页Journal of Naval Aeronautical and Astronautical University

基　　金：国家自然科学基金资助项目(61271444)。

摘　　要：航空磁探反潜作为航空反潜的重要手段,在其中发挥了重要作用。针对目前航空反潜作战中,磁干扰信号极大地影响对水下目标磁探测效果这个问题,文章先对输入信号进行预处理,并使用卷积神经网络实现对2种信号的识别。实验结果显示,卷积神经网络的方法对信号的识别率达到了85%,能够有效对信号进行准确地识别。As an important means of aviation anti-submarine,aviation magnetic anti-submarine played an important role.In current aviation anti-submarine warfare,the magnetic disturbance signal greatly affects the magnetic detection effect of underwater target.The article first preprocessed the input signal,and used the convolution neural network to recognize the two signals.The experimental results showed that the signal recognition rate of the convolutional neural network reached 85%,which could effectively recognize the signal accurately.

关 键 词：卷积神经网络 磁异常信号识别 干扰信号 

分 类 号：TN911.72[电子电信—通信与信息系统] TM153.1[电子电信—信息与通信工程]