Enhancement of frequency scanning interferometry signal for non-cooperative target based on generative adversarial network
Enhancement of frequency scanning interferometry signal for non-cooperative target based on generative adversarial network
复制标题
基于生成对抗网络的非合作目标频率扫描干涉测量信号增强
DOI：
10.1088/1361-6501/ac8c62
复制
发表时间：
2022-12-01
影响因子：
2.4
通讯作者：
Guo，Junkang
中科院分区：
文献类型：
--
作者：
Tian，Kai;Liu，Zhigang;Guo，Junkang
在非合作目标频率扫描干涉测量中，回波光功率低、干涉信号质量差、信噪比低。此外，频率扫描激光的使用带来的功率变化调制了干涉信号的幅度，并使幅度中心移位。传统的信号增强技术，如滤波，只能解决影响测量精度的部分问题，对这类信号进行全因子处理比较困难。提出了一种基于产生式对抗网络的非合作目标扫频干涉信号增强方法。通过学习样本数据集，可以在一定范围内提高信号的信噪比，同时对信号波形进行校正。仿真结果表明，该方法提高了非合作目标信号的信噪比，并对信号波形进行了较好的校正。最后，通过实验验证了该增强方法的有效性。
In non-cooperative target frequency scanning interferometry, the return optical power is low, the quality of the interferometric signal is poor, and the signal-to-noise ratio (SNR) is low. Moreover, the power change accompanying the use of the frequency scanning laser modulates the interferometric signal's amplitude and shifts the amplitude centre. Traditional signal enhancement techniques, such as filtering, can only solve some of the problems affecting the measurement accuracy, and the full-factor processing of such signals is difficult. This paper proposes a non-cooperative target frequency scanning interferometry signal enhancement method based on a generative adversarial network. By learning the sample dataset, the SNR of the signal can be improved within a certain range, and the signal waveform can be corrected simultaneously. The simulation results reveal that the SNR of the non-cooperative target signal is improved and the signal waveform is satisfactorily corrected. Finally, the effectiveness of the enhancement method was experimentally confirmed.