The Study Of Target Edge Imaging Method In Radio Tomographic Imaging

Target edge imaging is one of the key issues in the field of Radio tomography.It uses the imaging function of a radio tomographic imaging system based on the characteristics of wireless transmission channels to achieve accurate image reconstruction of the target edge in the scene.Because of the hardware limitations of the wireless communication transmission nodes,the errors of the linear model of the radio tomographic imaging system and the complexity of the wireless transmission channel,the target edge reconstruction in the radio tomographic image is inaccurate,the target image area is blurred,and the target streak aliasing is caused.The problem of insufficient accuracy of edge reconstruction makes it difficult to recognize and classify target images in the scene.After analyzing the linear imaging model of the radio tomographic imaging system,it is found that the accuracy of the reconstructed image is affected by the resolution of the reconstructed image,the linear transformation model of the reconstructed image,and the type of wireless channel characteristic parameters.This thesis focuses on the key techniques to improve the accuracy of target edge reconstruction in radio tomographic images,including radio tomographic image accuracy assessments,radio tomographic image super-resolution reconstruction methods,radio tomographic image degradation and recovery models,and channel status information-based radio tomographic imaging methods.The main contribution of this work is as follows.(1)Firstly,according to the composition of target image reconstruction algorithm in radio tomographic imaging system,the current actuality of target image reconstruction is summarized,and the accuracy of target image reconstruction of existing radio tomographic imaging algorithms is analyzed and compared.The current several linear imaging system models,linear weight models and image reconstruction algorithms are qualitatively analyzed.The current evaluation criterions of the target reconstruction accuracy are revisited in details,and the difference and correlation among them are pointed out,and the evaluation criteria of the subsequent chapters are confirmed.(2)In the existing radio tomographic imaging algorithms,the limitations of the total links of radio tomographic imaging system and the resources for image reconstruction operations cause low resolution of radio tomographic reconstruction images,resulting in insufficient target reconstruction accuracy in the radio tomographic reconstruction images.This thesis proposes a radio tomographic image enhancement algorithm based on image superresolution reconstruction for the requirement of high resolution images for radio tomographic reconstruction images.The algorithm uses the current radio tomographic imaging algorithm to generate low-resolution image sequences,and then uses the spatial domain image superresolution reconstruction algorithm to reconstruct the high-resolution image of the edge structure information in the low-resolution image.It improves the resolution and target details of radio tomographic images.Experiments show that the radio tomographic image enhancement algorithm based on image super-resolution reconstruction not only improves the resolution of radio tomographic images,but also can solve the problem of insufficient target edge accuracy in radio tomographic images.(3)In the process of constructing the linear imaging model of the radio tomographic imaging system,the error of the linear model causes the problem of blurring of the radio tomographic reconstruction image: the edge of the target edge is not clear,the target streak and the target area are heavily aliased,and the target area is difficult to be segmented.This thesis proposes a radio tomographic image target restoration algorithm based on the degraded function to solve the blur problem of reconstructed image caused by the linear model error.The algorithm firstly establishes the image degradation restoration model of radio tomographic imaging system,and mathematically expresses the blur problem of radio tomographic reconstruction images.Then the Gaussian mixture model is used to estimate the degradation function of the radio tomographic imaging system from the radio tomographic linear imaging model.Finally,a regularized image restoration algorithm is used to perform target image restoration processing on the radio tomographic image.Both simulation experiments and hardware experiments show that,the radio tomographic image restoration algorithm based on the degraded function can restore the target edge,and can reduce the aliasing effect of the target area and streak in the reconstructed image,and further improve the target edge imaging accuracy.(4)In indoor complex scenes,the multipath effect causes the received signal strength at the wireless receiving node to be unstable and fluctuates randomly within a large range.Radio tomographic imaging algorithms based on received signal strength cannot accurately reconstructe the scene,resulting in inaccurate edges of the target in the reconstructed image,and even false target regions that do not exist in the scene.The wireless channel state information,the sampling signal of the channel frequency response,has the ability to distinguish the signal multipath transmission characteristics,and can effectively cope with multipath interference problems in indoor complex scenes.This thesis proposes an indoor radio tomographic imaging algorithm based on channel state information.The algorithm first performs sub-carrier weighted average processing on the amplitude components of the channel state information to reduce the influence of multipath deep fading,and then performs linear transformation and singular value decomposition on the phase components of the channel state information to reduce the interference of multipath pair phase information.The feature quantities of the amplitude component and the phase component are obtained.Then,combined with the amplitude-weighted mean value and phase feature of the channel state information,the attenuation image on each pair of transmit-receive antennas is estimated.Finally,a radio tomographic image fusion algorithm based on contrast pyramid is used to fuse and reconstruct the tomographic image of the target in the scene from the attenuation images of multiple pairs of transmit-receive antennas.The experimental results show that the radio tomograhic imaging algorithm based on channel state information can reconstruct the edge information of the target in complex indoor environment,reduce the probability of false target regions in reconstruction image,and improve the accuracy of the target image.