By Wavelet based Fusion Directional Adaptive Interpolation Super Resolution Image Construction

Super resolution is the process of recovering a high resolution image from multiple low resolution images of the same scene. This process consists of three steps: Image registration, Image fusion and Interpolation. Image registration is the first step in super resolution process in which the reference and sensed images are geometrically aligned. This alignment is done based on maximization of an appropriate similarity measure. The similarity measure used in this paper is Pearson correlation coefficient. Methods used for fusion are Discrete Wavelet Transform (DWT) based decomposition at one level and DWT based decomposition at 2 level. Interpolation is the final step in super resolution process. Various interpolation algorithms namely, nearest neighbour, bi-cubic, edge oriented interpolation, angular variation based interpolation and directional adaptive interpolation were applied to the images and are compared quantitatively. The overall performance of super resolution algorithms are compared using sharpness index metric.