基于STFT谱图滑窗相消的微动杂波去除方法  

Micro-Doppler Clutter Removal Method Based on the Cancelation of Sliding STFT Spectrogram

作　　者：万显荣[1] 谢德强[1] 易建新[1] 胡仕波 童云 WAN Xianrong;XIE Deqiang;YI Jianxin;HU Shibo;TONG Yun(School of Electronic Information,Wuhan University,Wuhan 430072,China)

机构地区：[1]武汉大学电子信息学院,武汉430072

出　　处：《雷达学报（中英文）》2022年第5期794-804,共11页Journal of Radars

基　　金：国家自然科学基金(61931015,62071335);湖北省技术创新专项重大项目(2019AAA061);湖北省自然科学基金创新群体项目(2021CFA002)。

摘　　要：微动杂波往往具有较大的多普勒展宽,会抬高噪底、湮没弱目标,造成虚警和漏检。有效去除微动杂波对提高雷达性能具有重要意义。该文利用匀速目标回波和微动杂波在短时傅里叶变换(STFT)谱图中的形态差异,提出了一种基于STFT谱图滑窗相消的微动杂波去除方法。具体地,匀速运动目标回波在STFT谱图中表现为特定频率单元上平行于时间轴的直线型能量条带,而微动杂波具有时变非平稳特性,在STFT谱图中呈现出横跨多个频率单元的时变复杂形态。将原始STFT谱图沿时间维滑窗得到新的STFT谱图,则目标回波分布在这两种谱图中的相同位置,而微动杂波在这两种谱图中的位置存在明显差异。因此将上述两种谱图相减,根据相减前后谱图中各单元的强度变化情况,即可将目标回波和微动杂波分离,达到去除微动杂波的效果。仿真和实测结果均验证了所提方法的有效性。与常见基于时频变换的L-statistics算法相比,所提方法能够在去除微动杂波的同时,较好地保留了目标回波。Micro-motion clutter typically exhibits significant Doppler broadening,raises the noise floor,and annihilates weak targets,resulting in false alarms and missed detections.Removing micro-motion clutter effectively is critical to improving radar performance.In this study,a micro-motion clutter removal method based on the cancelation of the Short-Time Fourier Transform(STFT)spectrogram is proposed using the difference in the morphological performance of the constant-speed target echo and micro-motion clutter in the STFT spectrogram.The target echo appears in the STFT spectrogram as a linear energy strip parallel to the time axis on a specific frequency unit,whereas the micro-motion clutter appears as time-varying complex shapes across many frequency units due to its time-varying non-stationary characteristics.When the original STFT spectrogram slides along the time dimension to obtain the new STFT spectrograms,the target echo is distributed in the same position,whereas the position of the micro-motion clutter is different.Therefore,subtracting the above spectrograms,the target echo and the micro-motion clutter can be separated based on the intensity changes in each unit of the STFT spectrogram before and after subtraction,and the micro-motion clutter can be removed.The simulation and field experimental results validate the proposed method’s effectiveness.Compared with the common time-frequency-transform-based L-statistics algorithm,the proposed method can remove micro-motion clutter while retaining the target echo.

关 键 词：微多普勒效应 微动杂波 短时傅里叶变换 短时傅里叶变换谱图 

分 类 号：TN958[电子电信—信号与信息处理]