计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (16): 172-175.

• 图形、图像、模式识别 • 上一篇    下一篇

模糊Renyi熵与QGA结合的快速图像分割

赵 敏1,张 路1,孙棣华1,阳树洪2   

  1. 1.重庆大学 自动化学院,重庆 400030
    2.重庆大学 计算机学院,重庆 400030
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-06-01 发布日期:2011-06-01

Fast image segmentation based on fuzzy Renyi entropy and quantum genetic algorithm

ZHAO Min1,ZHANG Lu1,SUN Dihua1,YANG Shuhong2   

  1. 1.Department of Automation,Chongqing University,Chongqing 400030,China
    2.Department of Computer,Chongqing University,Chongqing 400030,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-06-01 Published:2011-06-01

摘要: 针对非模糊熵的阈值分割方法不能较好地反映数字图像本质上具有的模糊特性,提出一种新的基于模糊熵的图像阈值分割方法。通过模糊隶属度函数将图像直方图信息转换到模糊域,利用模糊Renyi熵计算目标与背景的信息熵。根据最大熵原理,引入量子遗传算法对隶属度函数参数进行寻优,进而得到图像的最佳分割阈值。与典型的阈值法进行对比实验,表明该方法能获得更好的分割结果,满足实时性需求。

关键词: 图像分割, 模糊Renyi熵, 量子遗传算法, 最大熵原则

Abstract: Digital image are by nature fuzzy,while traditional threshold method doesn’t reflect this fact.In order to retain the fuzziness of image,a new image threshold method based on fuzzy entropy is presented.Transforming the histogram of image into fuzzy domain by fuzzy membership function,the entropy of objects and background is calculated according to the definition of fuzzy Renyi Entropy.Following the maximum entropy principle,Quantum Genetic Algorithm(QGA) is employed to accelerate the search of the optimal parameters of membership function,and thus the best threshold of image is obtained by the combination of these parameters.The experimental results indicate that the proposed method obtains good performance,and satisfies the requirement of real-time.

Key words: image segmentation, fuzzy Renyi entropy, quantum genetic algorithm, maximum entropy principle