Research On Synthetic Aperture Radar Image Based On The Contourlet Transform

Synthetie Aperture Radar(SAR) has characteristic of all-weather,day-and-night,high-resolution and strong-transmission, so it is widely used in civilian and military fields.SAR is a coherent imaging system,so SAR images contain lots of Speckle noise. Contourlet transform is a new Multiresolution Geometry Analysis(MGA) method which can represent images most effectively. This paper uses Contourlet transform to SAR image processing,proposes new algorithms of speckle suppression and edge detection for SAR image and researches on abilities of these algorithms. The principal tasks are as follows.It studied the SAR image filtering methods in spatial domain and Contourlet transform. A new method for image de-noising which colligated the strong point of Contourlet transform and the magnitude of the detail of image in every scale and direction. This paper introduces gene m1 to adjust Donoho threshold and gene m2 to adjust multi-scale threshold using contourlet transform. Because mi < 1(i = 1,2), it solve the problem of snuffing out the contourlet coefficients to exceed, and mi relate to the levels and subbands, it also can solve the problem of taking no consideration of image detail. Experiment on image de-noising shows that: The denoised images obtained by these algorithms gain pretty good visual effect and all indices are enhanced effectively comparing to the other denoised images.These denoising algorithms protect target's texture and edge characters efficiently while suppressing speckles successfully.In this paper, the ROA and Duda edge detection methods of SAR image are introduced in detail. Due to the existence of speckle, classic image edge detection algorithm is not very suitable for SAR image. This paper gives a new effective algorithm which colligated the strong point of Contourlet transform and the Duda edge detector. First, based on the multi-direction and multi-scale Contourlet transform, we use the information of the high frequency in every direction to reconstruct, then use Duda edge detector to detect the edge in reconstructive image in every direction, at last sum up the information to obtain a final edge. This Algorithm can alse detect certain linear features in rapid speed. The experimental results show the effectiveness of the method.The tag of this paper makes a summary of research and gives expectation on future work。...