Through-the-wall Radar Imaging Using Compressive Sensing

Recently,through-the-wall-radar imaging(TWRI)has become a keen research,which has the capabilities of imaging and positioning in the area behind obstacles.Thus,it is widely used in antiterrorism,post-disaster rescue and extraction of life characteristics,etc.High resolution is an important indicator of radar imaging,which is generally achieved by two methods: one is increasing the bandwitdth of the transmitted signal,and the other is to increase the radar aperture.Traditional coherent imaging algorithm would produce a large amount of data and increase computational cost.Compressive sensing(CS)is applied to radar imaging is an emerging technology of imaging.Making use of the sparseness of the scene,it can achieve high-resolution images with limited sampled data.Based on the theory of CS,this paper carries out three aspects for TWRI:(1)Based on the theories of radar imaging and CS,the target echo is modeled and the sparse sampled signal is achieved.The Orthogonal Matching Pursuit(OMP),Bayesian Compressive Sensing(BCS)and Total Variation(TV)algorithms are taken as examples to analyze the reconstruction of imaging based on CS.Compared with the traditional coherent imaging algorithm,the radar imaging based on CS method reduces the amount of computation,resulting in low sidelobe and high resolution.The simulations analyze the advantages and disadvantages of three algorithms according to imaging quality and running time.It also analyzes the imaging quality under different sparsity and contrast,that verifies the performance of radar imaging based on CS.(2)Then,this paper deals with the issue of unknown wall parameter in TWRI.In the traditional CS imaging method,the wall parameter is known,and the dictionary matrix can be directly constructed and recovery the target signal.However,the wall parameter is unknown in practice.CS can not be directly applied to the sparse reconstruction.In this paper,an imaging method of parametric sparse representation is proposed with unknown wall parameter to obtain autofocus image.By constructing a parametric dictionary,the wall parameter is continuously corrected,in order to obtain optimal sparse representation of the target echoes.The method views the sparse reconstruction and the wall parameter estimation as optimal goals at the same time.Thus,the sparse reconstruction is a joint optimization problem.To solve the joint optimization problem,two kinds of methods are proposed in this paper: the first one is the image reconstruction based on global searching,which searches the wall parameter in a range of candidates.The image quality is evaluated by contrast.The other is alternative iteration.In the iterative framework,sparse reconstruction and wall parameter correction are alternated.The sparse reconstruction is obtained by orthogonal matching pursuit algorithm,and the wall parameter is corrected by the least square or the conjugate gradient.When the wall parameter is the best,an autofocus image is achieved.Two types methods proposed in this paper can obtain the wall parameters which is close to the true value in simulations.(3)Finally,the multipath ghost is studied in this paper.In the indoor environment where the wall is surrounded,electromagnetic waves not only propagate between the target and the antenna,but also propagate along other paths from the inner wall,floor and the ceiling,which produce multipath.These multipath signals cause false targets in the image,resulting in failure of the detection and identification of targets.Due to the theory of CS,each propagation path can be represented by an observation matrix.All matrices are assembled into a dictionary.The image is reconstructed is by CS recovery method.Within this method,the multipath signal is suppressed.On this basis,the gradient sparse is used to analyse the pieceswise-constant(block target).In this chapter,under the consideration of multipath suppression and block target,a multipath suppression method based on TV-CS is proposed.Simulation experiments show that this method can effectively suppress the multipath ghost in through-wall scene,which recovery the target contour and obtain high-resolution image.