华南理工大学学报(自然科学版) ›› 2011, Vol. 39 ›› Issue (2): 71-75,80.doi: 10.3969/j.issn.1000-565X.2011.02.012

• 电子、通信与自动控制 • 上一篇    下一篇

视网膜图像中视盘的快速自动定位方法

赵晓芳1 林土胜1 李碧2   

  1. 1.华南理工大学电子与信息学院,广东广州510640;2.广东外语外贸大学信息学院,广东广州510420
  • 收稿日期:2010-07-12 修回日期:2010-09-17 出版日期:2011-02-25 发布日期:2011-01-02
  • 通信作者: 赵晓芳(1982-),女,博士生,主要从事图像处理、模式识别研究 E-mail:aozhy119@126.com
  • 作者简介:赵晓芳(1982-),女,博士生,主要从事图像处理、模式识别研究
  • 基金资助:

    国家自然科学基金资助项目(60972136);广州市科技计划项目(2009J1一CA0I)

Fast Automatic Localization of Optic Disc in Retinal Images

Zhao Xiao-fang1  Lin Tu-sheng1  Li Bi2   

  1. 1. South China university of technology, electronic and information institute, guangdong guangzhou 510640; 2. Guangdong university of foreign studies institute of information, guangdong guangzhou 510420
  • Received:2010-07-12 Revised:2010-09-17 Online:2011-02-25 Published:2011-01-02
  • Contact: 赵晓芳(1982-),女,博士生,主要从事图像处理、模式识别研究 E-mail:aozhy119@126.com
  • About author:赵晓芳(1982-),女,博士生,主要从事图像处理、模式识别研究
  • Supported by:

    国家自然科学基金资助项目(60972136);广州市科技计划项目(2009J1一CA0I)

摘要: 基于眼底视网膜血管的分布结构及视盘本身的特点,提出一种快速自动定位视盘的方法.首先根据视网膜血管的网络分布结构大致定位视盘的垂直坐标;然后根据视盘的亮度信息及视盘与血管的关系来定位视盘的水平坐标;最后把视盘限定在以粗定位的视盘为中心的一个小窗口内,用Hough变换精确定位视盘中心.该方法不需要事先分割视网膜血管,也不需要对算法进行训练.实验结果表明,文中算法具有较高的定位精度和较快的定位速度.

关键词: 视网膜图像, 视盘, 坐标, Hough变换

Abstract:

Based on the structure of retinal blood vessels and the features of the optic disc(OD),a fast automatic method of optic disc localization is proposed.In this method,first,the vertical coordinate of the OD is determined according to the distribution structure of the retinal blood vessel network.Then,the horizontal coordinate is located based on the OD brightness information and the relationship between the OD and the blood vessels.Finally,the OD is limited in a small window with the rough location of the OD as the center,and the OD center is precisely positioned via the Hough transform.By this method,neither prior segmentation of retinal blood vessels nor algorithm training is required.Experimental results indicate that the proposed method is of high localization accuracy and speed.

Key words: retinal images, optic disc, coordinate, Hough transform