Research On MMW InSAR Imaging And Target Detecting With Sparse Sensing

Millimeter Wave Interferometric Synthetic Aperture Radiometer(MMW InSAR)imaging system receives millimeter wave radiation of target scene by an equivalent large aperture antenna synthesized by multiple small aperture antennas and obtains visibility function in the spatial frequency domain directly by operation of plurality correlation,ultimately gets millimeter wave brightness temperature images of target scene by inverting visibility function,which improves performance of millimeter wave radiation imaging system like spatial resolution and sensitivity effectively.It has the advantage of being better hidden,sensitive to metal target,not being affected by severe conditions such as dust,smoke,night.This system has been widely used in military and civilian areas like navigation,remote sensing,anti-terrorist and security check,and plays an important role in target detection area.The ultimate purpose of the InSAR imaging system is to obtain millimeter wave brightness temperature images with high quality and conducting effective target detection on them for different applications,so high performance reconstruction algorithm and target detection method are key component of InSAR imaging research.Based on sparse sensing theory,this thesis starts with fundamentals and models of InSAR imaging,further studies the features of imaging reconstruction and proposes new InSAR imaging reconstruction methods and target detection algorithm based on visual saliency theory.It mainly includes the following aspects:(1)Research on target properties,modeling and simulation,imaging reconstruction features of InSAR imaging:The radiation characteristics of millimeter-wave target are introduced.Modeling and simulation are conducted from a statistical property perspective.Relationship between plural correlation output of millimeter wave interferometry and millimeter wave bright temperature distribution in the far field region is established based on system transmission characteristics.This relationship is further extended to cover both the far field and the near field regions.The basic idea and method of InSAR imaging are presented and the problems of existing imaging reconstruction algorithms are analyzed.(2)Research on one-dimensional sparse imaging model and the corresponding one-dimensional imaging reconstruction algorithms:Starting from the perspective of numerical processing of InSAR imaging,statistical characteristics and specific prior information of millimeter wave InSAR images are analyzed and generalized,then an one-dimensional sparse imaging model of InSAR imaging is proposed based on Compressed Sensing(CS)theory.Also,sparse regularization imaging reconstruction algorithm is proposed for it,which reduces the amount of system data processing and improves the accuracy and robustness of the imaging reconstruction of InSAR.By further studying the sparse solution performance of sparse regularization reconstruction algorithm,reweighted L1 norm reconstruction algorithm is proposed adopting the weighted parameters improved by the statistical features of the millimeter wave InSAR images,which further improves the suppression ability of system noise and error interference and the accuracy of imaging reconstruction.(3)Research on two-dimensional sparse imaging model and the corresponding two-dimensional imaging reconstruction algorithms:Two-dimensional imaging model of InSAR imaging is obtained by modifying the one-dimensional imaging model and a two-dimensional sparse imaging model of InSAR is proposed based on low rank Matrix Completion(MC)theory.Two-dimensional singular value threshold reconstruction algorithm is proposed for this model,which reduces the amount of system data processing and overcomes the disadvantage of being difficult f to reconstruct large-dimensional millimeter wave brightness temperature images by using one-dimensional imaging reconstruction method.Based on unconstrained Lagrangian model,two-dimensional accelerated proximal gradient reconstruction algorithm is proposed,which further improves the speed of imaging reconstruction under low data processing load and the accuracy of imaging reconstruction.(4)Research on spectrum-based visual saliency detection algorithm for millimeter wave InSAR imaging with sparse sensing:According to the characteristics of InSAR imaging and the statistical characteristics of millimeter wave InSAR images,target detection model of InSAR imaging is established.Spectrum-based visual saliency detection algorithm for millimeter wave InSAR images is proposed.This algorithm removes the negative effects of information loss in the traditional detection methods based on space domain;enhances detection capability of weak targets and improves the speed and accuracy of detection.(5)The design and realization of millimeter wave InSAR imaging system:Based on binary interferometer,a scanning double-channel InSAR imaging system is realized,which performance is verified by the interferometric fringe test experiment.The estimation parameters of the imaging system are calibrated by the point auxiliary source and the imaging reconstruction experiment is carried out,which also validates the effectiveness of proposed imaging reconstruction methods based on sparse sensing.