基于统计模型的变分水平集SAR图像分割方法

针对SAR图像感兴趣区域分割问题，提出了一种基于统计模型的变分水平集分割方法。该方法在分析SAR图像特征的基础上，利用相干斑噪声的统计模型直接定义了关于水平集函数的能量泛函，不同于一般水平集方法中关于参数化曲线的能量泛函。通过极小化能量泛函，建立了水平集函数演化的偏微分方程。对水平集演化方程的数值求解，实现了对SAR图像感兴趣区域的分割。分别采用模拟和真实SAR图像对提出的方法进行了验证，试验结果表明该方法充分利用了SAR图像的特征信息，不需要相干斑噪声预处理，能够准确实现对SAR图像感兴趣区域的分割。

A Variational Level Set SAR Image Segmentation Approach Based on Statistical Model

In this paper, a variational level set approach for segmentation of Regions Of Interest (ROI) is presented. Following the analysis of the SAR image characteristic, a new energy functional is defined by importing the statistical model of speckle noise. The energy functional is with respect to level set function, which is obviously different from the energy functional with respect to parameterized curve in general level set approach. The segmentation of ROI is implemented by the numerical solution of the Partial Difference Equations (PDEs) that are derived by minimizing the energy formulation. The performance of the approach is verified by both synthetic and real SAR images. It shows that the approach has the advantage of importing the noise information in the energy definition, thus accurately extracting the ROI from SAR image but without any speckle pre-processing step.