基于PF-TBD的双基地微多普勒提取  

Particle Filter for Track-before-detect of Bistatic Micro-Doppler Signatures

作　　者：邹小海[1,2] 艾小锋[1,2] 李永祯[1,2] 王雪松[1,2] 肖顺平[1,2] 

机构地区：[1]国防科技大学电子科学与工程学院,长沙410073 [2]电子信息系统复杂电磁环境效应国家重点实验室,长沙410073

出　　处：《信号处理》2013年第3期304-312,共9页Journal of Signal Processing

基　　金：国家自然科学基金青年科学基金(61101180)资助课题

摘　　要：时频分布-Hough变换是目前单/双基地微多普勒参数化估计的主要方法。该文首先介绍了进动钝头双锥弹头双基地微多普勒模型,然后针对时频图中散射中心的幅值分布特性未知导致无法确定Hough变换的参数域空间的检测门限这一问题,该文引入了PF-TBD算法来检测散射中心存在性,思想是将时频图中散射中心微多普勒曲线视为目标的运动轨迹,利用PF-TBD算法实现对微多普勒曲线的检测前跟踪,再利用Hough变换提取微多普勒曲线,逐步检测和提取所有微多普勒曲线。最后针对某双基地下的进动钝头双锥弹头微多普勒时频图,提取了图中所有微多普勒曲线,验证了算法的可行性。TFD-Hough （Time Frequency Distribution-Hough） transform is the main method of estimating parameters based on the mono- and bistatic micro-Dopplers. Firstly, the micro-Doppler models of the bistatic scattering center of the precessing warhead are given. To solve the problem of detecting the scattering centers in the parameter domain of the Hough transform, the method of particle filter based track-before-detect （PF-TBD） algorithm for detecting the micro-Doppler curves is proposed. The micro-Doppler curve in the time frequency image are regarded as the target＇ s moving trace, then the PF-TBD algorithm is using to track and detect the micro-Doppler curves and the curves＇ parameters are extracted by Hough transform, all of the micro-Doppler curves are detected and extracted step by step. Finally, the time frequency image is obtained via the time-frequency analysis on the calculated bistatic scattering coefficients of the blunt-nosed biconieal war- head model, and the micro-Doppler curves are extracted from the time frequency image using the proposed method, so the feasibility of the method is validated

关 键 词：粒子滤波 检测前跟踪 微多普勒 时频分布 HOUGH变换 

分 类 号：TN957[电子电信—信号与信息处理]