应用膨胀算法提取目标的直接定位算法

摘要

为了解决被动雷达系统中的多发射源定位问题,提出了一种基于多重信号分类(MUSIC)算法和图像膨胀(IE)算法的直接定位方法.该方法结合了谱分析中的MUSIC思想,通过对接收量测协方差矩阵进行特征分析求解目标的位置.首先,在目标个数未知的前提下,利用Akaike信息准则(AIC)来确定模型阶数;然后,推导了基于MUSIC的定位代价函数;之后,利用图像膨胀算法处理得到的代价函数平面;最后,膨胀处理后的输出为目标个数及目标位置的估计值.提出的算法有效地解决了目标检测及提取的问题,能够确定多个目标的位置坐标,为后续的定位性能分析提供可能性,也保证了算法的完整性.进一步地分析了多个临近目标情况下影响目标提取性能的主要因素.%In order to solve the problem of localizing multiple transmitters in passive radar system,a direct position determination method based on the multiple signal classification(MUSIC)algorithm and the image expansion(IE)algorithm is proposed.This method combines the idea of MUSIC in spectral analysis and deter mines the positions of targets by performing feature analysis on the measured covariance matrix of the received signal.First,under the premise of the unknown number of the targets,the Akaike information criteria(AIC)is used to deter mine the order of the model.Then,the cost function for localization is deduced based on MUSIC.Afterwards,the IE algorithm is used to deal with the cost function plane.Finally,the outputs of the expansion processing are the estimated number and positions of targets.The proposed approach effectively solves the problem of target detection and extraction and can accurately locate multiple targets,which makes it possible to do the subsequent performance analysis and also ensure the completeness of the algorithm.Moreover,we analyze the main factors that affect the target extraction performance in the case where there are multiple adjacent targets.