消除halo效应和色彩失真的去雾算法-中国图象图形学报.DOC

中图法分类号：TP391.4 文献标识码：A 文章编号: 消除halo效应和色彩失真的去雾算法 刘兴云，戴声奎 华侨大学信息科学与工程学院，厦门 361021 摘 要：目的：雾天条件下采集的图像存在低对比度和低场景可见度的问题，传统的去雾算法易产生halo效应和色彩失真问题。为此，结合大气散射光特性提出一种基于相对总变差的图像复原方法。方法：首先从大气散射光与纹理信息无关的角度出发，利用相对总变差分离图像主结构和图像纹理信息准确估计大气耗散函数，通过引入一个自适应保护因子来避免复原图像的色彩失真问题，最后由大气散射模型计算复原图像并进行图像的亮度调整，得到一幅清晰无雾的图像。结果：通过与经典的去雾算法比较，表明该方法可以有效避免halo效应和天空颜色失真等不足，并且在图像的深度突变处也能得到很好的去雾效果。结论：实验表明该算法的场景适应能力较强，时间复杂度与图像的大小成线性关系，相比于前人的算法在计算速度上有一定的提高。 关键词：相对总变差；大气耗散函数；去雾；大气散射模型 A halo-free and color-distortion-free algorithm for image dehazing Liu Xingyun1, Dai Shengkui1 1.College of Information Science and Engineering , Huaqiao University , Xiamen 361021 , China Abstract：Objective: The foggy image has the problem of low contrast and low visibility, the traditional dehazing algorithms suffer from the halo phenomenon and color distortion in the dehazing image. Based on the characteristics of atmospheric veil, a image restoration algorithm based on relative total variation has been proposed in this paper. Method: Due to the atmospheric scattering light has nothing to do with the texture information, firstly, the atmospheric veil can be estimated accurately according to the main structure and texture information separating by relative total variation. Then a adaptive protection factor is used to avoid the distortion of restoring image. Finally, the ideal result is recovered through physical model and the brightness is adjusted by a curve. Result: Compared with the state-of-the-art dehazing methods, our method can avoid halo artifacts or color distortion and achieve good dehazing result at a distant scenes and places where depth changes abruptly. Conclusion: The experimental results show that the proposed method has a robust scene adaptability, and provide much faster computing rate due to the liner relation between time complexity and image size. Key words：relative total variation; atmospheric veil; dehazing; atmospheric scattering model 0 引 言 雾天环境下，大气中存在的粒子会与光线发生散射、折射、吸收等反应，使成像设备捕捉到的户外图像模