Research On Super-Resolution Reconstruction Method Of Infrared Image Sequences

In recent years,with the rapid development of infrared imaging technology,it has been widely used in both military field and civilian field,the requirement for increasingly strong infrared images of high quality.However,the density of infrared detector units is much lower than that of visible light image detector for the limitations of material and technological process,leading to low resolution of infrared images.It is the focus of research in the field of infrared imaging that how to improve the resolution of the infrared images based on the imaging system.So super-resolution image reconstruction algorithms emerge at the right moment,which aims to transcend the limiting resolution of imaging system by reconstruct high resolution image through the use of digital image processing,and it has been a key trend at present.Therefore,it is of great practical value to study super-resolution image reconstruction technology in the field of infrared imaging.This thesis focuses on the in-depth analysis methods of super-resolution image reconstruction.In multi-images reconstruction,this thesis especially explores and studies super-resolution reconstruction algorithms based on MAP method and POCS method.Firstly,this thesis introduces the implementation process and principles of the MAP algorithm,analyzing the reconstruction method based on markov random field model.This thesis proposes an adaptive threshold MAP reconstruction algorithm which combines the image gray level co-occurrence matrix.It has a strong ability to suppress noise and can better protect the image high frequency information.Secondly,this thesis analyzes the basic principle of the POCS algorithm.Aim at the questions of the traditional POCS method depended on the initial estimation of the high resolution image and the ringing artifact,a POCS algorithm based on edge and detail protection is proposed.The initial estimation of high resolution image is obtained through iterative curvature-based interpolation instead of bilinear interpolation,and correcting the point spread function to reduce the ringing artifact.Furthermore,the improved POCS algorithm can effectively enhance the image quality and edge effect.Finally,the simulation experiments and analyses utilizing infrared images are conducted through the algorithms of this thesis.The experimental results show that the proposed algorithms can acquire reconstruction infrared images with great visual effect and higher resolution.Moreover,the proposed algorithms can acquire reconstruction infrared images whose peak signal to noise ratio,root mean square error are improved,and structural similarity index is much higher.