基于改进C-V模型的高分辨SAR图像分割  

High-resolution SAR image segmentation based on improved C-V model

作　　者：王翠杰[1] 杨永红[1] 林明[1] 

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212003

出　　处：《电子设计工程》2017年第19期161-164,169,共5页Electronic Design Engineering

摘　　要：针对SAR图像灰度分布不均匀现象,提出利用对均匀和不均匀区域都能很好的拟合的G0概率密度函数对C-V水平集模型进行改进。针对经典的C-V水平集模型只利用区域信息而没有利用边缘信息,从而造成虚假边缘轮廓较多以及演化时容易陷入局部极小值等现象,提出在C-V模型中引入能够很好的去除SAR图像中的乘性噪声的改进的边缘检测函数,定位图像的边界以及控制曲线的演化速率,且增加能避免水平集函数重新初始化的距离正则项作为内能项。通过人工合成图像和真实高分辨SAR图像的分割实验验证了改进后的C-V水平集模型能够更快速度、更高准确度地定位目标边界,实现高分辨SAR图像分割。For the phenomenon of heterogeneous grayscale in SAR image, we use the G0 probability density function what can fit homogeneous and heterogeneous areas to improve the C-V model. For the classic C-V model only use the geographic information and don＇t use the edge information, so that it has many false profile of edge and easy to fall into local minimum, by using the edge detection function what can remove the multiplicative noise join in the C-V model, and a distance regularized term that can avoid the periodic updates of level set function become a internal force term. The experiments based on both synthetic and real high-resolution on SAR image prove the improved C-V model has faster division speed and accuracy.

关 键 词：边缘检测函数 G^0概率密度函数 距离正则项 SAR图像分割 C-V模型 

分 类 号：TN957.52[电子电信—信号与信息处理]