Super-Resolution Reconstruction of Radio Tomographic Image

To improve the resolution of radio tomographic image and get more details of the target, this paper introduces nonuniform interpolation, projection onto convex sets and structure-adaptive normalized convolution approaches for image reconstruction from a series of radio tomographic images. We compare these approaches in three aspects: time consumption, visual inspection and similarity calculation. Experimental results show that super-resolution reconstruction algorithms enhance the resolution of radio tomographic image while preserving the detail and edge of image, and the structure-adaptive normalized convolution approach is appropriate to radio tomographic imaging system.