基于短时迭代自适应-逆Radon变换的微多普勒提取方法

对于频率交叠严重且频率成分接近的多分量信号,常用的短时傅里叶变换(Short Time Fourier Transform,STFT)和S方法(S-Method,SM)频率分辨能力不足,重构精度低.针对该问题,本文结合逆Radon变换提出了基于短时迭代自适应-逆Radon变换(Short Time Iterative Adaptive Approach-Inverse Radon Transform,STIAA-IRT)的微多普勒特征提取方法.首先采用基于加权迭代自适应的STIAA时频分析方法分析了散射点模型的微多普勒特性,然后利用逆Radon变换分离重构不同散射点的微多普勒分量.该方法在低信噪比、邻近时频分布情况下能获得高分辨的多分量信号的完整微多普勒信息,性能分析显示STIAA-IRT重构精度较高,明显优于STFT-IRT (Short Time Fourier Transform-Inverse Radon Transform)和SM-IRT (S-Method-Inverse Radon Transform)特征提取方法.

Micro-Doppler Extraction Based on Short-time Iterative Adaptive Approach and Inverse Radon Transform

For multicomponent signals overlapping seriously and neighboring,the widely-used algorithms like the short-time Fourier transform ( STFT) and S-Method( SM) are poor in frequency resolution and low in reconstruction accuracy.To solve this problem,a method based on short-time iterative adaptive approach and inverse Radon transform ( STIAA-IRT) to extract the micro-Doppler signatures is proposed combining with the inverse Radon transform (IRT).Analysis of the micro-Doppler characteristics for a point scatterer model with the employment of the STIAA time-frequency transform based on weighted iterative adaptation is made,and then the IRT is utilized to separate and reconstruct different micro-Doppler compo-nents with the result that the complete micro-Doppler features of all useful signals are obtained successfully under low SNR and time-frequency distribution in high resolution.Finally,analysis of the performance illustrates that STIAA-IRT has a higher reconstruction accuracy along with an obvious advantage over the signature extracting algorithms of STFT-IRT and SM-IRT.