Research Of The Auto-focus Algorithms In Synthetic Aperture Radar

Synthetic aperture radar is one that bases itself on the principle of Doppler Frequency and the relating thesis on pulse coherent relations, which help SAR transcend the boundary of azimuth definition of the real antenna. Together with the technique of pulse compression, SAR manages to realize the forming of two-dimensional image of long distance. With its further development afterwards, more people took advantage of this technique to form image of high definition and imposed higher standards on the image. And for years, various auto-focus algorithms had been used to improve to reach people's high demands. In this paper, I have conducted a thorough research on different image-forming algorithms and various auto-focus algorithms, which includes theoretical analysis and relating simulations. On the basis of these work, I put out a brand new algorithm and verified its efficiency, which played an positive role in improving the SAR image's quality.This paper focused on the research on the SAR auto-focus algorithms. While the traditional ones are used mainly on the process of azimuth phase error (We can apply them into the range process as well), which is based on one prerequisite that the coupling of data between range and azimuth has already been removed, that is to say, range cell migration and range move have been perfectly corrected, the real situation does not work that way. Because of the uncertainty of the vector V (both quantity and direction), together with some rough procession, the application of the auto-focus algorithms can not reach the ideal point.Besides the thorough analysis and research of the two commonly used SAR image-forming algorithms (In fact, I have conducted respective simulation of the two), I have also conducted the theoretical computations (the comparison of Map Drift (MD) and Reflectivity Displacement Method (RMD) algorithms) and simulation of the Phase gratitude algorithms (PGA). By means of the study of the three kinds of traditional auto-focus algorithms including the model with or without parameters, auto-search of best parameters, I have an in-depth understanding of some important parameters and their respective roles in the whole image-forming process and got some meaningful insights viewing from different points in conducting various auto-focus algorithms.By means of expanding the concept of the traditional auto-focus, I put out the auto-adapting algorithm, which abandons the hypothesis that the range cell migration has been perfectly corrected which the traditional ones turn to. I applied the process to the stage of correction of range cell migration in the whole image-forming. The process helps to improve some important parameters'accuracies and can be used for further following processes.After enough theoretical analysis and mathematical computation to evaluate the efficiency of the self-adapting algorithm, I applied it into the process of real data within the framework of Chirp Scaling image-forming algorithm. On getting the results, I also conducted thorough analysis in the following three fields, internal set parameters'effect on the efficiency of the self-adapting algorithm, the effect which this algorithm played on the process of image-forming with wrong parameter and the algorithm's astringency. Through the comparison of real image from the raw data and analysis of data curves, I have reached a conclusion that this one has a more effect on the image quality as against traditional ones.In order to further improve the image quality and this self-adapting algorithms'efficiency, I also looked up relating information and practiced them in the real simulation and got some ideal results, which at the same time improve the algorithm itself.