Denoising Synthetic Aperture Radar Image Via Wavelet Packet Analysis

Synthetic aperture radar (SAR) emit electromagnetic waves to target through the active way,And using targets reflected the electromagnetic wave imaging. Synthetic aperture radar can obtained the high resolution image of the target at different frequency band and different polarization modes. It has characteristics of all-weather, day-and-night, high resolution, receiving not influence of cloud cover or illumination conditions. So it is widely used in space reconnaissance, aviation measurement, satellite oceanic observation, remote sensing, image matching guidance, etc. Therefore, SAR image processing technology research has important theoretical significance and application value.This paper reference a large amount of relevant literatures and have a deep research and analysis in the process of SAR imaging producing inherent noise of coherent imaging mechanism - coherent spot noise,and having research and analysis a variety of methods of the current inhibiting coherent spot noise,Including space domain filtering methods and wavelet transform domain filtering methods. This paper mainly according to SAR image noise characteristics put forward noise reduction method based on wavelet packet analysis. This paper research wavelet packet analysis application in the process of the SAR image noise reduction in detailed, and gives the wavelet packet threshold function, threshold, wavelet function and decomposition scale selection. Meanwhile in the noise reduction process ,soft and hard threshold eclectic method which plays a key role was improved. Through determining the thresholding function, threshold, wavelet function and the decompose scale eventually formed a complete a theory of SAR image denoising algorithm based on wavelet packet theory.Through experiment, the suggested method by the paper put forward,which in the process of elimination of SAR image noise make a ideal effect,and which Compared to the traditional filtering method has more flexibility and adaptive ability, Completely accords with this topic research purposes.