A super resolution algorithm based on L-R iteration geometric mean in electromagnetic imaging

There are mainly two reasons for the highly image blurring in the electromagnetic imaging system. On one hand, image is degraded by the filtering effect of the diffraction-limited system. On the other hand, under-sampling and noise cause the vague as well. The special resolution of the current electromagnetic imaging super-resolution algorithm cannot meet the requirement. Therefore, an algorithm combined with geometric mean interpolation and Lucy-Richardson iterative is proposed. Firstly, the degraded image is interpolated based on geometric mean of pixels so as to increase the image pixels and information. Then use LR iterative to build super-resolution for the image of the known point-spread function. Compared with the traditional algorithm, the special resolution of the recovery image is improved by 70% and 20% respectively under the condition of no-noise and 20dB noise reconstructed by the new method. At the same time, the algorithm has certain effects on noise suppression. Relevant simulations and experiments are practiced to check the correctness of the new algorithm.