首页|基于时频图像分类的信号识别方法研究

基于时频图像分类的信号识别方法研究

扫码查看
频谱管控旨在减少己方用频设备间自扰、互扰.掌握设备当前的用频情况,是对设备用频或辐射进行动态管控的前提.在实际应用中,往往难以保证用频设备上报信息的准确性,而监测设备采集到的设备信号也会呈现出相对复杂的样式.传统的信号识别方法主要针对固定、常发的信号,对于偶发、时变的信号难以发挥作用.提出了一种基于深度学习的信号识别方法,通过构造信号的时频图像,利用图像分类技术,提取信号动态频谱特征,提升对偶发、时变信号识别的准确率,通过仿真实验,验证了方法的有效性.
Research into Signal Recognition Method Based on Time-frequency Image Classification
Spectrum management and control aims to reduce self-interference and mutual interfer-ence between own frequency equipments.To master the current frequency usage of equipment is the premise of implementing dynamic control of equipment frequency usage and radiation.In prac-tical applications,it is often difficult to ensure the accuracy of the information reported through the frequency equipment,and the equipment signals collected by the monitoring equipment will also show a relatively complex style.The traditional signal recognition methods mainly target at the fixed and frequent signals,and it is difficult to play a role in the occasional and time-varying sig-nals.A signal recognition method based on deep learning is proposed.By constructing time-fre-quency images and using image classification technology,the dynamic spectrum features of signals are extracted to improve the recognition accuracy of occasional and time-varying signals,and the ef-fectiveness of the method are verified through simulation experiments.

spectrum monitoringsignal recognitiondeep learning

王文兵、窦雪倩、谢金池

展开 >

中国电子科技集团公司第二十九研究所,四川 成都 610036

频谱监测 信号识别 深度学习

2024

舰船电子对抗
中国船舶重工集团公司第723研究所

舰船电子对抗

影响因子:0.213
ISSN:1673-9167
年,卷(期):2024.47(6)