计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (15): 175-179.

• 图形、图像、模式识别 • 上一篇    下一篇

基于Retinex理论的小波域雾天图像增强方法

储昭辉1,汪荣贵2,方 帅2   

  1. 1.合肥学院 数学与物理系,合肥 230601
    2.合肥工业大学 计算机与信息学院,合肥 230009
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-05-21 发布日期:2011-05-21

Enhancement algorithm of misty image based on Retinex theory in wavelet domain

CHU Zhaohui1,WANG Ronggui2,FANG Shuai2   

  1. 1.Department of Mathematics and Physics,Hefei University,Hefei 230601,China
    2.School of Computer&Information,Hefei University of Technology,Hefei 230009,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-05-21 Published:2011-05-21

摘要: 针对传统的雾天图像增强方法在增强图像的轮廓和细节的同时不能有效抑制图像噪声这一问题,在Retinex理论基础上,提出了小波域图像亮度与噪声双估计模型的图像增强方法,利用小波变换多分辨率的特点,对小波分解后的低频系数进行波动抑制,对高频系数通过多子带独立系数模型进行亮度估计,同时使用不同的软-硬估计法对图像的噪声进行抑制。实验证明该方法不但能有效地提取图像的亮度信息,而且进行图像增强的同时能有效地抑制图像的噪声,且对于不同的雾天环境也有良好的自适应性。

关键词: 图像增强, Retinex算法, 小波变换, 多分辨率分析

Abstract: In view of the problem of the traditional enhancement algorithm can not suppress the noise effectively when the image’s details and outline are enhanced.A novel algorithm of estimating the image illumination and noise in the wavelet domain is proposed based on Retinex theory in this paper.By using the feature of the wavelet multi-resolution,to suppress the volatility of the low-frequency coefficient and use the model of multi-subband independent coefficient to estimate the high-frequency coefficient,at the same time using soft-hard estimate method to suppress the image noise.The experimental results show that the algorithm can extract the image illuminate vector effectively,but also can suppress the image noise and adapt in different foggy weather.

Key words: image enhancement, Retinex, wavelet transform, multi-resolution analysis