SAR Image Processing Based On 2-D Hybrid Transform And Filtering

Synthetic aperture radar (SAR) attracted extensive research interests due to the lots of advantages, and SAR images contain many ground targets information, make full use of these informations will play a tremendous role.However, the first, SAR image reflects the backscatter characteristics of radar wave, when radar transmitted the coherent signal on the target, the interfere effect between coherent signal and scatter signal will make images generated coherent spot noise, but speckle noise have lots of image informations; secondly, in practice, SAR system will inevitably suffer from various sources of interference, imaging performance will be deteriorated. For example, the echo signal of the anti-radar target will become very faint that cannot be detected. Especially in the electronic countermeasure, SAR system will be affected by the enemy's initiative oppressive electromagnetic interference and the quality of the reconstructed image will be badly affected. Applying the useful SAR information must be overcome and solve these two problems.This paper emphatically introduces how inhibit the coherent spot noise and extract useful information in interfere condition. (1) Based on the current image processing method, this paper innovatively proposes the post-process methods, and firstly applied in SAR image processing field, researches how to inhibit the speckle noise;(2) Based on the post-processing of SAR image, research the target echo signal, applying the 2-D hybrid transform in the denoising pre-processing for SAR target echo signal, and also research the performance of anti-interference;(3) Since the 2-D hybrid transform as a 2-D filter require huge calculation, propose 2-D recursive filtering method process SAR image and extract targets.Experimental data shows that mix DFT (DCT) and DWT in different directions make both advantages adequately using when treat complicated two-dimensional signal; 2-D recursive filter is applied to the condition that target signal in the airspace and frequency domain is mixed with noise, it could real-time process SAR images.These two kinds of methods made active contribution in improving the translatability of SAR images.