Synthetic Aperture Radar Data Compression Algorithm Performance Evaluation And Speckle Reduction Algorithm

As the development of Synthetic Aperture Radar(SAR), the data rate isbecoming very large. This has stimulated interest into methods of compressing boththe raw data and the imaging data. In order to select the appropriate compressingalgorithms, it is necessary to evaluate the performance of the compressingalgorithms. It is common practice in such studies to use only the mean squareerror(MSE) and visual inspection of the reconstructed imagery to determine theacceptability of a compression technique. In this dissertation, we suggest a numberof additional metrics and present a systematic evaluation environment which aids inthe interpretation of the type of the error introduced, and in the understanding of themechanism that produced the error. The evaluation system includes the evaluation ofthe effects of compressing both on SAR raw data and on SAR imagery.According to the analysis of the evaluation result, we can conclude that theSAR raw data compressing only has a little influence on the quality of SAR imageryCompared with the SAR raw data compressing, the SAR imagery compressing canresult in a relatively larger influence on the quality of SAR imagery.In this dissertation, the speckle suppressing algorithm is studied, whichincludes the multi-look processing and the multi-variate optimal speckle reduction inSAR imagery. Based on the practical results, we can conclude that the specklereduction in the single-polarimetric SAR imagery will result in the deterioration ofthe spatial resolution. Experimental results with simulated polarimetric SAR datademonstrate the effectiveness of several polarimetric speckle suppressing algorithm.In addition, the probability density functions for multi-look polarimetric signaturesare studied, and their influence on the performance of the speckle suppressingalgorithms is analyzed.