基于优化最小化框架的墙体成像算法

A Wall Imaging Algorithm Based on Majorization-Minimization Framework

在穿墙雷达建筑物布局成像中,针对现有成像算法因没有充分利用墙体本身的物理特性而出现墙体轮廓模糊、边缘不连贯以及成像过程耗时的问题,提出一种基于优化最小化框架的墙体成像算法。该算法首先利用像素块来表征墙体连续块状的物理特性,并将其引入信号模型,然后以LASSO(Least Absolute Shrinkage and Selection Operator)模型为基础,在优化最小化框架下构造稳健的优化目标函数,最后利用墙体回波信号的时移特性并结合卷积得到迭代过程的快速实现。实验结果表明,该算法对墙体成像特征明显,不仅保证了墙体轮廓特性,而且杂波少、分辨率高,并较大幅度减小了成像算法处理时间。

Due to the underutilization of physical properties of the wall,the existing sparse imaging algorithms have some problems in through-wall radar building layout imaging.For instance,the imaging of wall has obscure contour and discontinuous edges,and the process of imaging is very time consuming.This paper proposes a wall imaging algorithm based on majorization-minimization framework.Firstly,the continuous physical characteristics of the wall are characterized by pixelblock,which is introduced into the signal model.Then,on the basis of least absolute shrinkage and selection operator(LASSO)model,a robust optimization objective function is constructed under majorization-minimization(MM)framework.Finally,the corresponding iterative process is deduced by using time shift characteristic of the echo signal of wall and the convolution in time domain.The results show that this method guarantees the contour characteristics of the wall and also suppresses the clutter of the wall image,and furthermore,significantly saving the imaging time.