摘要
针对传统的的暗通道去雾算法天空区域颜色易失真,复杂度过高等问题,提出了一种避免颜色失真的快速去雾算法。首先,该算法利用图像中的像素点相对高度和亮度值来寻找合适的大气光值。其次,根据大气模型和暗通道原理,使用双三次插值算法和导向滤波结合解决复杂度过高的问题,再利用容差机制对其优化,得到去雾图像。实验结果表明,该算法不仅有效地解决了传统暗通道算法的天空区域失真问题,并且缩短了图像处理所需时间,实现了单幅图像高质量快速去雾。
Aiming at the problems of the traditional dark channel defogging algorithm, such as high complexity and easily deformable colors in the sky region, a fast defogging algorithm that avoids color distortion is proposed. First, the algorithm uses the relative height and brightness values of the pixels in the image to find the appropriate atmospheric light value. Secondly, according to the atmospheric model and the dark channel theory, using double three-difference algorithm and guide filtering to solve the problem of high complexity, and then using the tolerance mechanism to optimize it, get the image after defogging. Experimental results show that the algorithm not only effectively solves the sky region distortion problem of the traditional dark channel algorithm, but also reduces the time complexity and achieves high-quality and fast fog removal for a single image.
作者
张泽浩
周卫星
Zhang Zehao;Zhou Weixing(School of Physics and Telecommunication Engineering of China,South China Normal University,Guangzhou 510006,China)
出处
《电子技术（上海）》
2018年第10期52-56,共5页
Electronic Technology
基金
广东省科技项目(项目编号:2016A010101021)的资助
关键词
图像去雾
暗原色先验
大气光
双三次插值
容差机制
image defogging
dark color prior
atmospheric light
bicubic interpolation
tolerance mechanism
