基于HSV颜色模型的广义梯度矢量流图像分割方法  

GGVF Method for Image Segmentation Based on HSV

作　　者：王征[1] 章品正[1] 梁晓云[1] 周正东[1] 罗立民[1] 

机构地区：[1]东南大学生物科学与医学工程系,南京210096

出　　处：《数据采集与处理》2005年第4期394-397,共4页Journal of Data Acquisition and Processing

摘　　要：采用广义梯度矢量流(G enera lized grad ien t vector flow,GGVF)作为势能函数的Snake算法在处理灰度图像分割时具有较好的性能。但当处理彩色图像时,GGVF模型往往将彩色图像转化为灰度图像进行处理,此过程造成颜色信息丢失,因而容易受到阴影等伪边界灰度变化的干扰。本文提出了基于HSV颜色模型的GGVF,充分利用了图像的色彩信息进行图像分割。实验结果表明,该方法能够有效地区分彩色图像的真实边界与由阴影产生的伪边界,并取得较好的分割结果。The Snake algorithm uses the generalized gradient vector flow （GGVF） as the potential energy function and has a good performance when dealing with gray images. The conventional GGVF model cannot deal with the color image segmentation well, since the model is easily influenced by the false edge caused by the shadow. A new GGVF method based on HSV color model is proposed. It achieves the color image segmentation by using the color information. Experimental results show that the method can distinguish the real and the false edges, and obtain good segmentation results.

关 键 词：图像分割 SNAKE算法 HSV色彩模型 广义梯度矢量流 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]