摘要
针对现有去雾方法计算成本高、色彩失真及去雾不彻底等问题,提出一种结合瑞利散射理论与暗通道先验的单幅图像快速去雾算法。对瑞利散射理论进行研究,设计一种简单而有效的改进大气光估算方法;建立暗通道先验模型,对初始透射率图进行估计,采用线性高斯滤波器对初始透射率图进行平滑处理;通过大气散射模型恢复无雾图像。通过户外模糊图像的定性和计算复杂度实验,将实验结果与其它算法进行比较,验证了所提算法的处理速度明显优于其它算法。
To solve the problems of existing defogging methods of high computational cost,color distortion and incomplete defogging,a fast defogging algorithm for single image based on Rayleigh scattering theory and dark channel prior was proposed.Rayleigh scattering theory was studied,and a simple and effective method to estimate atmospheric light was designed.A dark channel prior model was established to estimate the initial transmittance graph,and a linear Gaussian filter was used to smooth the initial transmittance graph.The air scattering model was used to restore the fog-free image.Through qualitative and computational complexity experiments of outdoor fuzzy image,and comparing the experimental results with other algorithms,it is verified that the processing speed of the proposed algorithm is obviously higher than that of other algorithms.
作者
黄小芬
林丽群
HUANG Xiao-fen;LIN Li-qun(Department of Information Technology,Concord University College Fujian Normal University,Fuzhou 350117,China;College of Physics and Information Engineering,Fuzhou University,Fuzhou 350117,China)
出处
《计算机工程与设计》
北大核心
2021年第9期2585-2591,共7页
Computer Engineering and Design
基金
国家自然科学基金项目(61801119)
福建省中青年教师教育科研基金项目(JAT160668)
福建省本科高校教育教学改革研究一般基金项目(FBJG20180130)。
关键词
图像去雾
大气散射模型
瑞利散射
暗通道
高斯滤波
image defogging
atmospheric scattering model
Rayleigh scattering
dark channel
Gaussian filter
