[attachment=15229]
[attachment=15230]
[attachment=15231]
[attachment=15232]
[attachment=15233]
1. Introduction
There is always a demand for higher quality and higher resolution images. The level of image detail is crucial for the performance of many computer vision algorithms. Current imaging devices typically consist of arrays of light detectors. A detector determines pixel intensity values depending upon the amount of light detected from its assigned area in the scene. The spatial resolution of images produced is proportional to the density of the detector array: the greater the number of pixels in the image, the higher the spatial resolution. In many applications, however, the imaging sensors have poor resolution output. When resolution can not be improved by replacing sensors, either because of cost or hardware physical limits, one can resort to resolution enhancement algorithms. Even when superior equipment is available, such algorithms provide an inexpensive alternative. This problem of recovering a high quality HR image from a set of distorted (e.g., warped, blurred, noisy) and LR images is known as super-resolution. A sequence of images of the same scene taken from slightly different positions(a small non-deliberate motion of camera between clicks would be enough) or taken at slightly different times would have similar but not identical information. This implies an increase in sampling rate as compared to a single image. Processing these images we can obtain an image of higher bandwith than any of the original images. As part of our project we have implemented a very popular super-resolution algorithm called the Iterative Back- Projection (IBP) Algorithm.This algorithm was proposed by Irani et, al [1] in the year 1990 but is still widely used to achieve efficient and effective super resolution. It is assumed that the imaging process is known and the relative shifts of the input pictures are calculated as precisely as possible using an image registration algorithm. An initial guess for the high resolution image is made and the imaging process simulated to obtain a set of simulated low resolution images. An error function between the actual and simulated low resolution images is defined which is minimized iteratively until no further improvement is obtained, or until the maximum number of allowed iterations is reached.