J4 ›› 2012, Vol. 39 ›› Issue (6): 1-9.doi: 10.3969/j.issn.1001-2400.2012.06.001

• 研究论文 •    下一篇

一种简化PCNN模型在彩色图像边缘检测上的应用

邵晓鹏;钟宬;王杨;黄远辉   

  1. (西安电子科技大学 技术物理学院,陕西 西安  710071)
  • 收稿日期:2011-07-18 出版日期:2012-12-20 发布日期:2013-01-17
  • 通讯作者: 邵晓鹏
  • 作者简介:邵晓鹏(1973-),男,教授,博士,E-mail: xpshao@xidian.edu.cn.
  • 基金资助:

    国家部委预研基金资助项目(9140A0106110DZ0125);中央高校基本科研业务费资助项目(JY10000905012)

Application of a simplified PCNN model in color image edge detection

SHAO Xiaopeng;ZHONG Cheng;WANG Yang;HUANG Yuanhui   

  1. (School of Technical Physics, Xidian Univ., Xi'an  710071, China)
  • Received:2011-07-18 Online:2012-12-20 Published:2013-01-17
  • Contact: SHAO Xiaopeng

摘要:

提出了一种改进的彩色图像边缘检测方法来克服传统方法不考虑色度信息及噪声影响而产生漏检、错检边缘的不足.通过提取图像的颜色主轴来综合表示图像的亮度和色度信息,并将彩色图像降维成包含色度信息的灰度图像用以检测; 为了降低噪声对检测结果的影响,采用脉冲耦合神经网络(PCNN)模型.由于PCNN模型中的参数过多,不利于控制,故使用简化的PCNN模型来减少参数,达到比较好的控制.实验表明,这种基于颜色主轴的PCNN彩色图像边缘检测方法不仅能准确得到彩色图像的边缘信息,而且对噪声有很强的抑制作用.

关键词: 边缘检测, 图像处理, 颜色主轴, 脉冲耦合神经网络

Abstract:

An improved color image edge detection method is proposed to overcome the defects of miss-detection and false-detection in traditional methods which are caused by thinking little of chroma information and the influence of noise. This paper exploits the color principal axis to represent both the luminance and the chroma information so that the color image can be converted into a pseudo gray image where color information is contained. Besides, we apply the Pulse Coupled Neural Networks (PCNN) model to reduce the influence of noise. Since we expect fewer arguments in the PCNN model to make our adjusting easier, a simplified version of PCNN is used to solve this problem. Experiment proves that this method which takes advantage of both PCNN model and color principal axis method can help us obtain the exact edge of the color image while the bad influence of noise is significantly suppressed.

Key words: edge detection, image processing, color principal axis, pulse coupled neural networks (PCNN) algorithm

中图分类号: 

  • TN911.73
Baidu
map