In the paper the most efficient model to implement super resolution using discrete wavelet transform is shown. The model is a three step process of image registration, interpolation and noise filtering using DWT. The paper resolutions the deduction of noise in the digital gray scale images. Normally, Data, text, picture, can be tainted by an additive noise with the process of the scanning. This method precludes the different type of noise such as Salt and Pepper noise (SP Noise) that reasons black and white spots in the original image. All the process is explained and simulation result is presented to prove the theory.