首页 >  2009, Vol. 13, Issue (2) : 224-231

摘要

全文摘要次数: 4099 全文下载次数: 3169
引用本文:

DOI:

10.11834/jrs.20090242

收稿日期:

修改日期:

PDF Free   HTML   EndNote   BibTeX
一种用于SAR图像分割的几何活动轮廓模型
国防科技大学,电子科学与工程学院信息工程系 湖南 长沙 410073
摘要:

基于SAR图像边缘检测算子和变分水平集方法,依据能量最小化准则,提出了一种新颖的几何活动轮廓模型.其基本思想是,直接定义关于水平集函数的能量泛函,并将经典模型中基于梯度算子的边缘指示函数,替换为基于ROEWA算子的边缘指示函数,提高了模型对于SAR图像的边缘检测能力;同时在新模型中加入水平集函数惩罚项,确保水平集函数逼近符号距离函数.由于该项的作用,模型的数值求解可采用简单的显式差分格式迭代,并保持较快的收敛速度.针对仿真图像、Radarsat和华东电子研究所实测数据的实验结果表明,该模型具有实现简单、分割边界定位准确和收敛速度较快等优点.

A geometric active contour model for SAR image segmentation
Abstract:

The geometric active contourmodel is a classical image segmentationmodelbased on the curve evolution theory and the level setmethod, which has been successfully applied to the segmentation ofmedical images. Due to the existence of speckle noise,themodel fails in SAR image segmentation. Moreover, there are severaldisadvantageswith thismode.l Firs,t the evolution equation\nisn’tobtained with the energy minimizationmethod. Second, the level set function needs to be reinitialized to a signed distance function periodically during the evolution. Finally, themodel is computationally inefficien.t Based on SAR image edge detectors and the variational level setmethod, a novelgeometric active contourmodel is proposed under the criterion ofenergyminimization. The basic idea is that the energy functional is defined directly on the level set function and the original edge indicator function based on gradients is replacedwith a new edge indicator function based on the ROEWA operator. Thus, the ability ofdetecting edges and the accuracy of locating edges are greatly increased, whichmakes themodelvery appropriate forSAR image segmentation. In addition,a term penalizing the level setfunction is added to the energy functional in order to force the level setfunction to be close to a signed distance function and therefore completely eliminates the need of the costly re-initialization procedure. Thanks to the contribution of this term, the numerical calculation of themodel can be implemented by a simple explicitdifference scheme; at the same time the evolution speed keeps very fas.t The proposedmodelhas severaladvantages.Forexample, itcan be easily implemented; itresults in accurate segmentation boundaries; it converges fast and its level set function doesn’t need to be reinitialized. The experimental results on the simulated image and realdata show its efficiency and accuracy.

本文暂时没有被引用!

欢迎关注学报微信

遥感学报交流群