Super-Resolution Image Reconstruction Based on Guidance Image

Super-resolution framework uses multiple noisy low resolution images from camera to generate a higher resolution image that has better spatial resolution than any of the available low resolution images. Super-resolution is an ill-posed problem. A inherent difficulty is the challenge of inverting the image observation model without amplifying the effect of noise in the measured data. Classically, the issue is addressed by incorporating regularization in the cost function to constraint the space of solutions. The main focus of this paper is to develop a regularization function that would preserve edges with improved resolution of super-resolved image. The proposed method is also compared with the classical superresolution methods and experimental results show the effectiveness and robustness of this method both visually and quantitatively.