论文部分内容阅读
提出了一种基于复小波域的自适应贝叶斯估计SAR图像相干斑抑制算法。该方法将传统小波的软阈值去加性噪声的特性与双树复小波变换的方向选择性结合起来,可以自适应选择有效的阈值来保留边缘等细节。应用贝叶斯估计来计算复小波系数的最佳值。该估计算子假设信号和噪声的复小波变换系数分别满足阿尔法稳定分布和高斯分布。实验结果表明该算法的去噪性能优于现有的基于离散小波变换的算法。
An adaptive Bayesian estimation speckle reduction algorithm based on complex wavelet domain is proposed. This method combines the characteristic of soft threshold de-additive noise of traditional wavelet and the direction selectivity of double-tree complex wavelet transform, and can select effective thresholds adaptively to preserve the details such as edges. Apply Bayesian estimation to calculate the best value of complex wavelet coefficients. The estimator assumes that the complex wavelet transform coefficients of the signal and the noise satisfy the alpha stable distribution and the Gaussian distribution, respectively. Experimental results show that this algorithm is better than the existing algorithms based on discrete wavelet transform.