计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (11): 172-174.DOI: 10.3778/j.issn.1002-8331.2009.11.052

• 图形、图像、模式识别 • 上一篇    下一篇

分形特征模糊增强及其在目标检测中的应用

张坤华,杨 烜,张 力   

  1. 深圳大学 信息工程学院,广东 深圳 518060
  • 收稿日期:2008-12-17 修回日期:2009-02-13 出版日期:2009-04-11 发布日期:2009-04-11
  • 通讯作者: 张坤华

Fractal feature enhancement based on fuzzy sets and its application in target detection

ZHANG Kun-hua,YANG Xuan,ZHANG Li   

  1. College of Information Engineering,Shenzhen University,Shenzhen,Guangdong 518060,China
  • Received:2008-12-17 Revised:2009-02-13 Online:2009-04-11 Published:2009-04-11
  • Contact: ZHANG Kun-hua

摘要: 提出一种新的运用模糊集理论进行分形特征增强的方法,并将其运用到复杂背景下的目标检测中。首先根据分形的尺度不变性计算图像各像素分形特征,定义新的隶属度函数对分形灰度图进行模糊特征平面映射,再运用非线性变换的模糊增强运算提高目标和背景的分形差异。在此基础上,结合数学形态学理论提出了基于增强分形特征的目标检测方法。实验证明该算法能够确实有效的提高目标和背景的分形特征差异,并克服了传统模糊增强算法的弊病;基于增强分形特征的检测方法,保证了复杂背景下目标检测的正确性和可靠性,并具有较好的抑噪性能。

关键词: 分形特征, 模糊集, 目标检测, 模糊增强

Abstract: Detecting target under complex background through single fractal feature would bring errors because of non-universality and fuzziness of fractal feature between target and background.A novel algorithm of fractal feature enhancement based on fuzzy sets is proposed in this paper.First,the fractal feature according to fractal scale invariance is estimated.Then,the new membership function of fuzzy is defined to map the fractal gray image into fuzzy feature plane.Finally,the fractal difference between target and background is increased by fuzzy enhancement transformation.On this basis,combing with mathematical morphology,the target detection method based on the enhanced fractal feature is given.The experimental results indicate this fuzzy enhancement algorithm can enhance the fractal feature of target and background effectively,and that the disadvantages of traditional fuzzy enhancement methods are avoided.Through detection method based on enhanced fractal feature,the target can be detected in complex background correctly and reliably,and the capability of restraining noise is good in detection.

Key words: fractal feature, fuzzy sets, target detection, fuzzy enhancement