提出一种基于自适应超完备字典学习的SAR图像降噪。该算法建立在超完备字典稀疏表示基础上,具有较强的数据稀疏性和稳健的建模假设。算法依据相干斑噪声统计特性,通过分步优化字典原子和变换系数自适应构造超完备字典,利用获得的超完备字典将图像局部信息投影到高维空间中,实现图像的稀疏表示,运用正则化方法建立多目标优化模型。最后通过对优化问题的求解重建SAR图像场景分辨单元的平均强度,实现SAR图像的降噪。实验结果表明,该算法对相干斑噪声有很好的抑制效果,并且具有保持图像细节信息的优点。

In this paper,a de-speckling algorithm for SAR images using an adaptive over-complete learning dictionary is proposed.The algorithm is based on sparse representation of SAR images via an over-complete dictionary It has strong data sparseness and provides solid modeling assumptions for data sets.First,a practical optimization strategy based on statistical properties of the speckle noise is used to design a redundant dictionary via an iterative loop.Second,the SAR image is projected into a high dimensional space using the learning dictionary and a sparse representation of the SAR image is obtained.Third,a model for multi-objective optimization problem is built by a regulation method.Finally,the de-noising process is realized through a solution of the multi-objective optimization problem in which the mean backscatter power is reconstructed.The experimental results demonstrate that the proposed algorithm has good de-speckling capability while preserving image details.