论文部分内容阅读
针对传统各向异性扩散抑斑算法存在的均匀区域噪声平滑不充分、边缘随迭代弱化及迭代次数的确定缺乏理论指导等问题,提出了一种新的各向异性扩散抑斑算法,该算法采用信息论匀质性测度作为图像中匀质区域与边缘的鉴别因子,使扩散系数能够更准确地控制扩散强度与扩散速率,从而达到充分平滑均匀区域噪声及保护边缘的目的。基于各向异性扩散方程求解与鲁棒误差范数最小化的等效性,提出了一种各向异性扩散方程的迭代停止准则。利用合成孔径雷达图像对本文算法的抑斑和边缘保持性能进行了仿真实验验证。结果表明,本文算法在均匀区域相干斑噪声抑制、边缘保持等方面均取得了优于传统算法的效果。
Aiming at the problem that the noise in the uniform region of the traditional anisotropic diffusion speckle smoother is not smooth enough, the edge weakens with the iteration and the number of iterations is lacking in theoretical guidance, a new anisotropic diffusion speckle reduction algorithm is proposed. The algorithm uses Information homogeneity measure, as discriminating factor of homogeneous area and edge in image, can make diffusion coefficient control diffusive intensity and diffusion rate more accurately, so as to achieve the purpose of smoothing even area noise and protecting edges. Based on the solution of the anisotropic diffusion equation and the equivalence of minimizing the robust error norm, an iterative stopping criterion of anisotropic diffusion equation is proposed. Synthetic aperture radar (SAR) images are used to verify the performance of the algorithm and the edge preserving performance. The results show that the proposed algorithm outperforms the traditional algorithm in suppressing the speckle noise in the uniform region and edge preserving.