Research On High-Resolution Imaging Method Combined With SAR And Full-Wave Electromagnetic Field Inversion

Microwave imaging has been widely used driven by the needs of biomedical imaging,non-destructive testing,through-wall and ground-penetrating measurement,human security inspection and so on.But higher imaging resolution and faster imaging speed have always been the focus pursued in this field.Microwave imaging originated from synthetic aperture radar(SAR)technology in the military field.It is essentially a kind of electromagnetic inverse scattering problem.The electromagnetic properties of the target area are deduced inversely through scattered electromagnetic field.There are many algorithms in the field of microwave imaging which have the advantages of fast imaging speed and wide range of application scenarios,such as back-projection(BP)algorithm and range migration(RM)algorithm.However,the imaging resolution of these algorithms is limited by the wavelength,which is difficult to further improve.The Gauss-Newton inversion(GNI)method can reconstruct the electromagnetic properties of the target area with high precision.However,the computational complexity of inversion algorithm increases significantly with the increase of target area and imaging resolution.Studies have shown that providing prior information on the shape or material of the target area for the inversion algorithm can effectively improve imaging resolution and efficiency.Currently,there are some studies using computed tomography(CT)or magnetic resonance imaging(MRI)techniques to provide shape prior information for inversion algorithms to reduce computational complexity.However,these methods cannot be implemented by microwave imaging systems and have disadvantages such as low price-performance ratio and poor flexibility.Based on solving the electromagnetic inverse problem,an imaging method which combines SAR technology and GNI method is proposed in this paper.Firstly,the BP algorithm is used for fast pre-imaging and supplying prior information about the target area,so as to reduce the size of the reconstruction area and assign appropriate initial values.Then GNI algorithm with multiplicative regularization(MR)is used to reconstruct the spatial distribution of the relative permittivity of the target area.On the premise of retaining the ultra-high resolution of the GNI algorithm,this method takes advantages of the fast imaging speed of the BP algorithm to make up for the disadvantages of the GNI algorithm in terms of operation speed and memory usage.The design is carried out from an overall perspective,and the entire system is built on the microwave system.The existing technology is mature,the equipment is costeffective,and the operation requirements are low.Only the same set of microwave measurement equipment can be used to meet the measurement requirements of BP algorithm and GNI algorithm at the same time.The acquisition of prior information does not require additional measurement instruments and operation links which leads to high engineering feasibility.In this paper,a simulation model is established based on the common situations of biomedical imaging,non-destructive testing and geological survey.According to the designed process,the MATLAB programming simulation is carried out.And the difference between the imaging resolution and speed of MR-GNI method in the two cases with and without prior information is compared.It is demonstrated that the use of prior information improves the realtime and accuracy of imaging.The simulation of the influence of antenna position error and initial value of scatterers is carried out.And the results show that the imaging resolution and speed of the method proposed in this paper depend on correct positioning and initial value close to the actual value.However,these errors are not concealed and are easy to be found and corrected.