The Research On The Algorithms Of Super-resolution Remote Sensing Image Reconstruction In Spatial Domain

Super-resolution(SR)image reconstruction is the technology of reconstructing a frame of image with high resolution from a group of warped,blurred and noised low-resolution(LR)images or video sequence about the same scene.The goals of this technology are to break through the resolution limitation of the imaging hardware facilities,to make up the loss of spatial resolution during the acquisition and transmission of images by the help of the data fusion of complementary information between multi-frame images,and then to reconstruct images with higher resolution.It can not only improve the visual effect of image,but also be convenient and beneficial for computer to analyze,process,recognize and understand the image.So far,the technology of Super-resolution image fusion has been extensively applied to remote sensing,military,public security,computer vision,medical imaging,multi-media electronic consuming,image compression and so on.Therefore,this technology has great prospect with significant applicational value.This thesis firstly introduces the developmental process of Super-resolution image fusion as well as all kinds of algorithms basing on both frequency domain and spatial domain. Secondly this thesis sets up the observation model,and then provides a detailed introduction to the principle,classification,and manipulation of image registration.This thesis adopts a gradient image registration method which is based on the Taylor theory.Since it can provide a good registration for big shifted image.Thirdly,this thesis focuses on the study of the principle,realization and steps of the POCS(Projection onto Convex Sets)methods.And a lot of experiments on the times of circle,threshold to analyze the factors influence POCS reconstruction result.In the light of the weakness of the enlargement of the end vibration and nosise of the traditional method,this thesis update the projection operator of the imaging system.So this thesis adds relaxation projection operator to restrain the enlargement of the end vibration and nosise.The efficiency of this algorithm can be recognized from the experimental results which are attached at the end of this thesis.Fourthly,This thesis focuses on the study of the Maximum A Posteriori(MAP)method with the introduction of principle,mathematical description and different prior knowledge of image model.Then it indicates the prior knowledge of image model Gausse-Markov which was adopted in the experiment.Due to the computing complexity of the MAP method,this thesis use gradient descent method to solve the algorithm.The experimental results show that the MAP algorithm can be able to efficiently reconstruct a high quality image.This thesis ends with the study of the Non-uniform Spatial Interpolation method with the introduction of principle.This thesis put forward a new method Based Non-uniform Distance interpolation.From principle,model,manipulation and steps of the development to introduce the new method.The advantage of the algorithm can be recognized as simple,compution simpleness.However, no optical fuzzy,motion-blurred and the prior knowledge of image in the method.