Image Super Resolution using Fourier-Wavelet transform

The low resolution images taken from a scene may contain crucial information that are barely visible to the eye. Super Resolution is the process of combining multiple noisy, blurry, low resolution images into a high quality, high resolution image. By registration, we fuse images taken at different times, at different angles of the same scene. Restoration and denoising of the fused images play a key role in Super Resolution. The multiframe Super Resolution algorithm applied here is MForWarD. It is a fast two step algorithm. First, Fourier-based Weiner filtering produces a sharp but noisy image. The next step uses Wavelet based denoising to remove noise artifacts. The algorithm is applied on several test images including remote sensing images and the results are presented.