基于改进二维Ostu算法的SAR图像分割方法研究

Research on the SAR Image Segmentation Method Based on Improved Two-Dimension Ostu Arithmetic

摘要:

INS/SAR组合导航系统对SAR图像分割提出了较高的实时性和准确性要求。文章针对SAR图像的特点与二维Ostu方法在区域划分以及运行速度上的缺陷,提出了一种改进的二维Ostu方法,更为准确地划分了目标和背景区域,并且将阈值搜索空间由二维降为一维。经仿真验证,该方法具有良好的分割效果并能有效提高分割速度,满足了组合导航系统相应的设计要求。

Abstract:

INS/SAR navigation system requires SAR image segmentation to be fast and precise. Aiming at the feature of the SAR images and the defect of two-dimension Ostu method in area partition and performing rate, an improved two-dimension Ostu method was put forward. The method partitions the objective area and background area precisely and reduces the threshold domain from two-dimension to one-dimension. Being proved via experiments, improved two-dimension Ostu method in the text act well.