The Research On Imagesuper-resolution Reconstruction Algorithm

Image super-resolution reconstruction (SRR) refers to a resolution enhancementtechnology, which reconstruct the image of high-resolution (HR) image from a set ofdegraded low-resolution (LR) images.This technique doesn’t need to change theexisting imaging equipment, just uses the signal processing to improve the imageresolution. It has very important application value and broad prospects inremote-sensing, business,medicine, video, security monitoring and other fields.In this paper, Firstly, the mathematical model of SRR is introduced, and severalclassical SRR algorithms are discussed. Frequency algorithms and spatial algorithmsare the two main reconstruction technique. Now, iterative back projection (IBP)approach, projections onto convex sets (POCS) approach，maximum a posterioriprobability （MAP）approach are wide used in SRR algorithms. Secondly, LR imagemust be inserted into the HR image grid, so image interpolation theory isinvestigated.Polynomial interpolation contain nearest neighbor interpolation, bi1inearinterpolation,cubic interpolation, B-Spline interpolation.This paper carry out research on SRR of the Single frame image and present anew reconstruction algorithm of SRR based on separate ideas.Afterwards，the paperderive and analyze the cascade template algorithms in the process of algorithmimplementation.Simulation results conferm the low-complexity and high-availabilityof the new algorithm.In order to improve the effect of SRR, we present the B-Splineinterpolation of multi-frame images based on Sub-pixel displacement.The paper leadB-spline interpolation technology into SRR of multi-frame images，and then use theadditional information of multi-frame images where the sub-pixel displacementexisted to improve the resolution. Simulation results show better visual effects andbetter performance with the new algorithm, furthermore， the information ofinterpolated image closer to the original image.