基于小波包分解的SAR图像斑点噪声抑制

Speckle Restraint of SAR Image Based on Wavelet Packets Decomposition

论述利用小波包分解技术和软门限理论对合成孔径雷达(SAR)图像斑点噪声(speckle)进行抑制与滤除.首先对SAR图像进行小波包分解,选取最优基,加大小波系数之间的差距;然后应用Donoho的软门限(softthreshold)理论,针对SAR图像斑点噪声的特点,在小波包变换域内选取合适的门限进行滤波;最后,通过小波包重构技术获得去除斑点噪声的SAR图像.实验证明,该方法能有效抑制SAR图像中的斑点噪声.

This paper describes an algorithm of speckle restraint for SAR image using wavelet packets decomposition and soft threshold theory. First, the SAR image is decomposed by wavelet packets,and the optimized wavelet bases are chosen in order to increase the difference between wavelet coefficients.Then,Donoho soft threshold theory is applied to restrain speckle of SAR image and a suitable threshold is selected. Last,SAR image after despeckling is reconstructed using wavelet packets reconstruction technique. The experimentation shows that the speckle is effectively restrained using the algorithm.