Wang Zhenghuan

Image reconstruction algorithms for radio tomographic imaging

Radio tomgographic imaging (RTI) is a method of imaging the attenuation caused by objects obstruction within the area surrounded by wireless sensor nodes. The image reconstruction of RTI is an ill-posed problem because the pixels are more than the RSS measurements. Therefore, conventional methods such as least square method are not applicable because the solutions amplify the noise by dividing small singular values. In this paper, we discuss three algorithms i.e. LBP, Tikhonov regularization and projected Landweber iteration to obtain the imaging results. LBP is very simple and needs little hardware resources. Tikhonov regularization method makes a good balance between image quality and regularization error and we use L-curve method to choose regularization parameter. Projected Landweber iteration method can further improve the imaging results while it needs lots of computation. All the three algorithms can be used to meet different requirement of applications. The experiment results are also presented to validate the effectiveness of algorithms.