Interpolation And Its Application In Super-resolution Reconstruction

SRC Super Resolution),which has an another name HR(High Resolution), is an important technique in image processing.The idea behind super-resolution imaging is to combine the information from a set of slightly different low resolution images of the same scene and use it to construct a higher resolution image. The differences between the low resolution images can have many origins, such as camera motion or change of focus.Some regular image reconstruction methods can reconstruct a high resolution image only within its cut-off frequency and have to lose the information beyond the frequency of the image. SR is the technique trying to avoid losing any of the information of image. It is one of the most active fields in the world of image processing recently. SR (which can be seen a inverse process of optic) reconstruct super (high) resolution image of one or set by using low resolution images (which have strong relationship between each other) without noise and aliasing.Interpolation is very important in the process of SR image reconstruction because of reducing noise and aliasing, but different methods also can bring cost of system computation and different affections to the result.In this paper we introduce interpolation and SR image reconstruction based on a model of one frequency SR reconstruction method.We first analysis the categories of interpolation and the difference between traditional and new methods. Interpolation is very important from the angle of image sampling. The affection of interpolation is showed by experiments.Secondly, we illustrate the art of state of super-resolution in the world and analysis of practicality of super-resolution. Based on the characteristic, the good and shortcomings of SR is analyzed.At last, we use a frequency domain technique to precisely register a set of aliased images, based on their low-frequency, aliasing-free part. After registration, with all the information of the set of low resolution, a super resolution image is then reconstructed using interpolation which is second phase of SR image reconstruction through values of different points in the low resolution images. Then we analyze affections of different interpolations to the reconstructed image.