The space target image super-resolution reconstruction technology research
|School||PLA Information Engineering University|
|Course||Photogrammetry and Remote Sensing|
|Keywords||Space Target Super-resolution Point spread function Image restoration Blind deconvolution Support domain De-noising|
As more and more platforms into space , more and more people pay attention for the study of the space of the target image . According to the the space target imaging results , make full use of technical methods of image super - resolution reconstruction of space technology to generate close to or even more than the optical diffraction limit of the target image , is clearly a good way . This thesis focuses on the the space target image super-resolution reconstruction of research and practice, the main work of this paper are as follows : 1, analysis of the basic principles of the super-resolution reconstruction intuitive way , for the first time in more detail the possibility of super-resolution reconstruction ; 2, a mathematical model of the image blind deconvolution , nature and the existing methods , in-depth analysis and iterative blind deconvolution (Iterative Blind Deconvolution, IBD) , maximum likelihood ( maximum Likelihood , ML ) method , Richard - Lucy ( RL ) method with non-negative and support the domain constraints recursive inverse filtering (Nonnegativity and Support Constrants Recursive Inverse Filtering, NAS-RIF) algorithm , done a lot of testing , compare various algorithms performance. 3, a sequence program using the maximum likelihood method dealing with spot image . This method does not use the reference star data from a number of noisy spot frame reconstruction extended target . Blindly determine the image support domain , target image thresholding approach is proposed to determine the target support domain , and achieved good results . 5 , based on the analysis of the classic median filtering technology , an adaptive weighted adaptive weighted denoising filter method , and achieved satisfactory results .