Research On Inverse Algorithms For Targets Buried In Heterogeneous Medium With Unknown Parameters

Target inversion in heterogeneous medium belongs to the inverse scattering problems. The inverse scattering problems mean that estimating the properties and position information of the targets with the received field data(acoustic wave or electromagnetic wave). Target inversion and imaging in heterogeneous medium has widespread applications, such as tumor target detection in heterogeneous biological tissues, and subsurface target detection and so on. In real world problems, the background medium where the target buried in usually has a complex unknown heterogeneous distribution. In the medium which has unknown heterogeneous distribution of acoustic or dielectric parameters, how to accurately reconstruct the target in heterogeneous medium is the main research contents of this dissertation. This dissertation proposes some new inverse algorithms for reconstructing the targets concealed in the medium with structural heterogeneity and the multilayered media with unknown properties.In order to accurately reconstruct the targets buried in the medium with structural heterogeneity, a new iterative reconstruction method(IRM) is proposed and then applied in the microwave induced thermo-acoustic tomography(MITAT). The IRM combines the time reversal mirror(TRM), fast marching method(FMM) and simultaneous algebraic reconstruction technique(SART) to iteratively update the medium with structural heterogeneity. The IRM can accurately image the targets buried in the heterogeneous medium with unknown properties by solely using the measured data. In this dissertation,the anatomically realistic numerical breast phantoms are used to test the feasibility and efficiency of the IRM. Moreover, the data experimentally measured from the MITAT system are also employed to validate the performance of the proposed IRM. Through the simulated and experimental results, the proposed IRM can efficiently reduce the effects of structural heterogeneity in the biological tissues and therefore accurately reconstruct the tumor target and increase the image quality.Aiming to reduce the computation amount and computing time for modeling the medium with structural heterogeneity, an iterative reconstruction method with the optimal time-of-flights(TOFs) is proposed and then applied in the MITAT. The proposed method adaptively selects the optimal TOFs according to the changing trends of TOFs with and without the heterogeneity. With the selected optimal TOFs, the parameters of the assumed homogeneous medium are iteratively updated by using the TRM, FMM and SART. The updated model of the heterogeneous medium is subsequently employed to reconstruct the concealed targets. Through the simulations and MITAT experiments, the image quality reconstructed by the proposed optimization method is nearly the same to that reconstructed by the traditional IRM while the proposed optimization algorithm reconstructs the model of the heterogeneous medium faster than the traditional IRM.To accurately detect the target buried in layered medium, an electromagnetic inverse scattering series method(EISSM) is proposed and applied in 2-D and 3-D problems. The advantage of the EISSM is that it can position the buried target accurately without knowing the dielectric parameters of layered background medium a priori. Both numerical simulations and experiments are performed to test the proposed EISSM. From the simulated and experimental results, the EISSM can locate the buried target with less position errors than the existing method such as TRM.In order to accurately reconstruct the targets buried in layered media with high dielectric properties, a modified EISSM is proposed and applied in the 2-D problems. The position errors predicted by the modified EISSM and the traditional EISSM are also analyzed. The modified EISSM introduces a phase correction factor to compensate the error of the correction term calculated by the traditional EISSM. Both the simulations and the experiments demonstrate that the modified EISSM can efficiently reduce the position error predicted by the traditional EISSM, and therefore accurately reconstruct the targets buried in layered media with high dielectric properties.In this dissertation, we thoroughly investigate the problems of the target inversion in heterogeneous medium with unknown parameters. Then the targeted studies based on the target concealed in the medium with structural heterogeneity and the multilayered media are carefully performed. Both numerical simulations and experiments are employed to demonstrate the capability of the proposed methods. Moreover, the influence factors and limitations of the proposed methods are also discussed. This work has provided a reference for further research in the heterogeneous medium with unknown parameters.