华南理工大学学报(自然科学版) ›› 2010, Vol. 38 ›› Issue (2): 132-136,148.doi: 10.3969/j.issn.1000-565X.2010.02.025

• 电子、通信与自动控制 • 上一篇    下一篇

基于双重弯曲度积的图像拐点检测

廖文志 皮佑国   

  1. 华南理工大学 自动化科学与工程学院, 广东 广州 510640
  • 收稿日期:2009-01-20 修回日期:2009-04-18 出版日期:2010-02-25 发布日期:2010-02-25
  • 通信作者: 皮佑国(1953-),男,教授,博士生导师,主要从事图像处理与模式识别研究.E-mail:auygpi@scut.edu.cn E-mail:mrlwz@163.com
  • 作者简介:廖文志(1982-),男,博士生,主要从事图像处理与模式识别研究.
  • 基金资助:

    粤港关键领域重点突破招标项目(20071684);广东省工业攻关项目(2005B10201003)

Corner Detection of Image Based on Product of Two Bending Degrees

Liao Wen-zhi   Pi You-guo   

  1.  School of Automation Science and Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China
  • Received:2009-01-20 Revised:2009-04-18 Online:2010-02-25 Published:2010-02-25
  • Contact: 皮佑国(1953-),男,教授,博士生导师,主要从事图像处理与模式识别研究.E-mail:auygpi@scut.edu.cn E-mail:mrlwz@163.com
  • About author:廖文志(1982-),男,博士生,主要从事图像处理与模式识别研究.
  • Supported by:

    粤港关键领域重点突破招标项目(20071684);广东省工业攻关项目(2005B10201003)

摘要: 针对拐点处与非拐点处的弯曲度值差别小的问题,文中提出了一种基于双重弯曲度积的图像拐点检测算法,通过第一重弯曲度与第二重弯曲度的乘积扩大拐点处与非拐点处的弯曲度值的差别,并对数字图像进行拐点检测实验.结果表明,该算法不仅简单有效,而且具有较高的稳定性和准确性,能够解决实际的图像拐点检测问题.

关键词: 拐点检测, 弯曲度, 图像拐点, 曲率

Abstract:

As the difference in bending degree between comer points and non-corner points is not large enough for image detection, an algorithm of image comer detection based on the product of two bending degrees is proposed. The proposed algorithm enlarges the differences in bending degree by multiplying the first bending degree and the second bending degree. Then, the comer detection for digital images is conducted. Experimental results show that the proposed algorithm is simple and effective, and that it helps to detect image comers with good stability and high precision.

Key words: corner detection, bending degree, image corner, curvature