基于饱和度运算的快速图像去雾算法  

A fast Image Dehazing Algorithm Based on Saturation Operation

作　　者：凡遵林 毕笃彦[1] 李权合[1] 马时平[1] 

机构地区：[1]空军工程大学航空航天工程学院,陕西西安710038

出　　处：《计算机仿真》2014年第12期251-255,共5页Computer Simulation

基　　金：国家自然科学基金"基于视觉认知的大景深雾霾退化图像去雾技术研究"(61372167)

摘　　要：在雾、霾等恶劣天气条件下,大气介质中悬浮粒子的散射和吸收作用会严重退化户外拍摄的图像,造成图像识别率降低。从单色大气散射模型出发,提出面向视觉感知的HSV颜色模型定义饱和度的新算法,估计大气光亮度A;利用饱和度运算估计透射率函数进而恢复场景反照率;最后经直方图拉伸得到最终的复原图像。实验结果表明,改进算法能有效地复原场景的对比度和颜色,提高图像的视觉度和计算效率。Bad weather, such as fog, haze and mist, can significantly degrade the visibility of outdoor scenes due to the suspending particles in the atmospheric medium. This algorithm starts with the atmosphere scattering model, defines a novel saturation operation based on the HSV color model for visual perception, and estimates the atmosphere lightness A; then computes the transmission map by using saturation operation in order to recover the scene albedo; finally uses histogram stretching to get the dehazing image. Results on a number of outdoor foggy images demonstrate that the proposed method achieves good restoration for contrast and color fidelity and improves image visibility and computational efficiency greatly.

关 键 词：图像去雾 大气散射模型 饱和度运算 直方图拉伸 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]