论文部分内容阅读
根据SAR图像的特性,提出了一种基于全变差最小化的Bandelet变换域SAR图像相干斑抑制方法。该方法首先对SAR图像进行Bandelet变换,得到图像的Bandelet系数和几何流,然后采用主成分分析法提取图像的Bandelet系数和几何流的主成分,利用全变差最小化的方法确定阈值,并通过循环递归运算处理重构后的图像。实验表明:该方法在抑制图像相干斑的同时较好地保留了几何结构信息,并去除了伪吉布斯现象。
According to the characteristics of SAR images, a method for speckle reduction of Bandelet transform based on total variation minimization is proposed. Firstly, the Bandelet transform of SAR images is performed to obtain the Bandelet coefficient and geometric flow of the image. Then the principal component analysis (PCA) is used to extract the Bandelet coefficients and the principal components of the geometric flow, and the threshold is determined by the method of total variation minimization. Recursive loop processing of the reconstructed image. Experimental results show that the proposed method preserves the geometric information and suppresses the false Gibbs phenomenon while suppressing the speckle of the image.