Research Of Improving Resolution Of SAR Images Based On Regularization

Synthetic Aperture Radar(SAR) is a kind of imaging radar, which is base onthe microwave remote sensing, and it has strong penetrability during all-day andall-weather. It is of great importance and application prospect in not only militaryfield, but also other sectors, such as the national economy. Now, multi-band,multi-polarization, variable beam and variable angle are the ways that the realSAR is heading. The quality of SAR images determine the efficiency of SARapplication, The higher quality of SAR images, the better its performance after theapplication become.(such as target recognition, feature extraction, imagesegmentation). Lots of application domains ask SAR image has very highresolution, so to study and discuss how we can improve the resolution of SARimage has very important and far-reaching sense.There are four basic methods of how to improve the resolution of SARimage, and among them, this paper will make a deep research on regularizationmethod. The specific work of this paper can be summarized as follows:Firstly,This paper introduces the model for the observation of SAR imageclearly, and explains the significance of why we have to improve the resolution ofSAR image. On the same time, the sparse distribution of SAR image is analyzed,a standard for judging the quality of SAR image is given.Secondly,The basic algorithm of regularization method is improved, inventsa new way to handle SAR image, using te requirement for regularization methodby the two dimensional decomposition. So the two dimensional decomposition ofthe SAR image is possible. In the process of iterative solution, some study andimprove base onl knorm regularization method and iterative solution is studied.And all these method has been proved to be right.Lastly,Starting from the basis of sparse and compression perception, andattacking it from different angles as variation model and Bayesian theory, this paper studies use the regularization functional model in order to enhance theresolution of single channel SAR image. This paper uses the Lorentzian functionas a sparsity constraint term, estimates and solves the parameters of the model.The stability, the compressibility and the generalized sparse of the solution in theprocess of image processing is analyzed. This paper uses spaceborne SAR imageto perform the data validation, which proves the conclusion of this paper is veryeffective.