基于改进正则化方法的SAR图像增强技术

SAR Image Enhancement Using Modified Regularization Method

常规SAR成像算法分辨率受系统带宽限制 ,并存在严重的相干斑干扰 ,不利于提取图像特征 .正则化方法通过施加先验信息的约束 ,可以获得稳健的重建图像 .由于SAR图像动态范围非常大 ,直接利用正则化方法重建图像 ,图像的细节特征不明显 .本文采用改进正则化方法 ,以图像模值的对数变换作为势函数变量 。

Conventional SAR imaging techniques suffer from resolution and speckle noise, which complicate feature extraction from image. Regularization technique could be used to obtain robust reconstruction image by imposing prior information. Direct regularization does not help much to emphasize details due to drastically large range of SAR image. A modified regularization algorithm is considered to take logarithm transformation on norm of image. Experimental result demonstrates the effectiveness.