基于正则模型的SAR图像自适应目标特征增强方法

Adaptive Threshold Selection for SAR Image Enhancement Based on Regularized Model

研究SAR图像特征增强的自适应阈值选取方法。本文采用一种新的度量函数表示SAR图像的稀疏先验,建立正则化模型,证明了不动点迭代求解的收敛性。然后将SAR相位历史数据转化为复数据,得到复数域正则解的解析表示式。该方法不需要迭代,简化了求解过程,并且将正则化参数的确定归结为阈值的选择问题。最后基于广义交叉检验准则实现了阈值的自适应选取。实验中采用目标杂波比来衡量处理效果,实验结果说明本文方法能快速有效地实现SAR图像特征增强。

Adaptive threshold selection for SAR image enhancement is studied. Using a new measure function to express the sparse prior information of SAR image, a regularized model is established. The model is solved by the fixed point iteration method and proved to be convergent. Then, transforming the data from phase history to the complex value, it can obtain a closed-form solution of the complex domain regularized model. The simplified method has no need to use the iteration and considers the problem of determining regularization parameter as threshold selection. Finally, the optimum threshold is adaptively selected based on the generalized cross validation （GCV） technique. Experimental results measured by the target-to-clutter ratio （TCR） demonstrate that the proposed method can fast and effectively enhance SAR image.