A Computationally Efficient Super-Resolution Reconstruction Algorithm Based On The Hybird Interpolation

Abstract

In order to release the limitations on the computational complexity and the detail missing of the traditional interpolation algorithm, this paper proposes a double channel hybrid interpolation algorithm for super-resolution reconstruction that can not only reduce the complexity effectively but also keep the details of the image sufficiently. The improved algorithm involves following methods: first, the parallel processing is used in this project; second, the cubic B-spline linear interpolation is introduced in the low frequency domain; third, the edge interpolation algorithm is applied in the high frequency details. The simulation and experiments both indicate that this algorithm is able to process the image in real time with details preserved.