摘要
针对SAR图像特殊的噪声特性,提出一种基于灰色模糊熵的快速图像分割方法。该方法不仅考虑了图像像素的灰度信息,还考虑了像素与其邻域像素的空间相关信息。为弥补传统模糊方法对噪声敏感的缺陷,引入灰色关联分析理论,设计图像当前像素灰度值与其八邻域像素灰度值组成的比较序列,通过计算其与目标点参考序列的灰色关联度修正传统隶属函数,以更精确地描述该灰度值属于目标或背景的模糊隶属度,并进一步给出了灰色模糊熵模型作为选取最佳阈值的准则。此外,为尽快确定最佳隶属度阈值,采用了具有群体智能的粒子群优化算法。实验表明该方法可以在一定程度上抵制SAR图像噪声,快速得到较为清晰的分割结果。
Aiming at the special character of noise in SAR image, this paper presented a fast SAR image segmentation method based on grey fuzzy entropy, in which not only utilized the gray level information of each pixel, but also involved its spatial correlation information within the eight neighborhood pixels. To decrease the noise-sensibility of the traditional fuzzy entropy methods, it introduced the theory of grey relational analysis to modify the existing fuzzy membership function by selecting the referential sequence and the compared sequences from the original image, and then employed the grey correlation degree of the two kinds of sequences to better the membership function. Consequently, designed a grey fuzzy entropy function to locate the best segmenting threshold. In addition, used PSO, as an swarm intelligent tool, to speed up the segmenting procedure. Some experimental results indicate that the method not only ignores the disturbance of inherent speckle in SAR image, but also provides with some better segmented objects.
出处
《计算机应用研究》
CSCD
北大核心
2009年第10期3968-3970,共3页
Application Research of Computers
基金
国家自然科学基金资助项目(60803088)
陕西省自然科学基金资助项目(2007D07)
中国博士后科学基金资助项目(20060401009)
关键词
合成孔径雷达
图像分割
灰色关联分析
模糊熵
粒子群优化
synthetic aperture radar （SAR）
image segmentation
grey relational analysis
fuzzy entropy
particle swarm optimization （ PSO ）
