一种改进的单幅图像去雾算法  

An improved algorithm for single image defogging

作　　者：田小平[1] 崔倩男 吴成茂[1] 

机构地区：[1]西安邮电大学电子工程学院,陕西西安710121

出　　处：《西安邮电大学学报》2017年第2期25-31,共7页Journal of Xi’an University of Posts and Telecommunications

基　　金：国家自然科学基金重点项目(61136002);陕西省自然科学基金资助项目(2014JM8331;2014JQ5183;2014JM8307);陕西省教育厅科学研究计划资助项目(2015JK1654)

摘　　要：为提高雾天室外拍摄图像的质量,给出一种单幅图像去雾增强算法。引入大气耗散函数,建立一种简化的大气散射模型;根据雾霭在图像中与距离所呈现出的衰减特性,通过构造指数模型,获取更精确的大气耗散函数;根据大气散射模型实现对雾天图像的复原。实验结果表明该算法能够适应雾霭浓度变化,并恢复真实场景的色彩和清晰度,提高图像质量。An algorithm for single image defogging is proposed to improve the quality of images token outdoor in foggy scenes. A simplified atmospheric scattering model is established by introducing the atmospheric dissipation function. According to the attenuation characteristics of fog whose density decreases inversely with the distance, a more accurate atmospheric scattering function is set up by constructing an exponential model. By using the extablished model, foggy images can be restorated more clearly. Experimental results show that, the proposed algorithm can adapt to some mist concentration, recover the scene color and clarity, and improve image quality.

关 键 词：去雾 图像增强 大气散射模型 大气耗散函数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]