摘 要: 針對傳統Retinex算法采用高斯濾波估計圖像的照射分量易產生邊緣模糊,不能有效去除脈沖噪聲且處理后的圖像顏色易失真等問題,提出一種基于三邊濾波的Retinex圖像去霧算法。該算法利用三邊濾波器估計圖像的照射分量,三邊濾波器繼承了雙邊濾波器既可以有效降低圖像加性高斯噪聲又可以保持圖像邊緣細節的特性,同時又解決了雙邊濾波器與高斯濾波器不能有效濾除脈沖噪聲,易產生偽邊緣等問題。為驗證該算法的有效性,采用5種不同的客觀評價參數對處理后的圖像進行評價。實驗證明,該算法能有效地改善霧天圖像的退化現象,提高圖像的清晰度。
關鍵詞: Retinex算法; 圖像去霧; 三邊濾波器; 雙邊濾波器
中圖分類號: TN919?34; TP391 文獻標識碼: A 文章編號: 1004?373X(2013)24?0077?04
Retinex image defogging algorithm based on trilateral filtering
CAO Yong?mei, ZHANG You?sai
(School of Electronics and Information, Jiangsu University of Science and Technology, Zhenjiang 212003, China)
Abstract: A Retinex image defogging algorithm based on trilateral filtering is proposed in this paper to avoid edge fuzziness, impulse noise and color distortion in traditional Retinex algorithm. In the new algorithm, the trilateral filter is adopted to estimate the illumination component on image. The trilateral filter is utilized to replace Gaussian filter and bilateral filter, both of which can not effectively filter the pulse noise but are easy to produce 1 edge. Trilateral filter can preserve the image’s edges while it suppresses the additive white Gaussian noise as the bilateral filter does. Five different objective evaluation parameters are used to evaluate the disposed images to prove the effectiveness of the algorithm proposed in this paper. Experiment results show that this algorithm can effectively improve the degradation of foggy images and enhance their definition.
Keywords: Retinex algorithm; image defogging; trilateral filter; bilateral filter
0 引 言
近年來,隨著計算機視覺領域的發展,霧天圖像的清晰化問題逐漸成為該領域的研究熱點。目前對于圖像去霧技術的研究大體上可以分為兩類[1]:基于大氣散射物理模型的圖像去霧方法和基于圖像增強的去霧方法?;诖髿馍⑸湮锢砟P偷膱D像去霧方法從造成圖像退化的具體原因角度分析,側重于考慮大氣散射作用對成像過程的影響,對霧天圖像的退化過程進行數學建模,求解模型參數,再通過圖像退化的逆過程來恢復原來的真實圖像?;趫D像增強的去霧方法不考慮圖像退化的原因,而是有選擇性的突出圖像中感興趣的特征,減弱不需要的特征。一般通過增強圖像的對比度和校正圖像的顏色以達到圖像去霧的目的,從而使處理后的圖像更適合人類的視覺特性或計算機識別系統。……