融合区域和边界信息的水平集SAR图像分割方法

提出了一种基于区域和边界信息的水平集SAR图像分割方法。该方法根据SAR图像的区域统计特征和边界梯度信息,建立SAR图像分割能量泛函模型;通过最小化能量泛函得到曲线演化偏微分方程;采用水平集方法求解演化方程,实现了SAR图像的分割。分别采用模拟和真实SAR图像对该方法进行了仿真。实验结果表明,该方法能充分利用SAR图像特征,不需要去除相干斑噪声的预处理过程,实现了对图像中目标与背景的正确分割。

A Level Set SAR Image Segmentation Method with Combined Region and Edge Information

In this paper, a new level set synthetic aperture radar (SAR) image segmentation approach based on region and edge information is proposed. An energy functional which is adapted for SAR image segmentation is defined. The energy functional consists of a region-based term derived from maximum-likelihood estimation of a mixed Gamma model and a boundary-based term derived from geodesic active contour model. Partial differential equations (PDEs) of curve evolution are obtained by minimization of the energy functional. To implement image segmentation, the solution of the PDEs by a level set approach is proposed. The efficiency of the method is verified by both synthetic and real SAR images. Experimental results implement the more accurate and rapid SAR images segmentation without preprocessing steps to filter speckle noise.