摘要
雾霾天气导致成像设备获得的图像质量下降,影响了人类视觉感知及计算机视觉特征提取。计算机视觉系统对去雾的速度要求较高,因此,提出一种快速图像去雾算法。首先根据雾天图像退化模型,通过高斯模糊对环境光进行估计,其次根据图像整体亮度对大气光值进行估计,最后根据雾天图像退化模型复原无雾图像。实验结果证明,在主观视觉效果方面,去雾效果明显,明显改善了图像质量。在去雾速度方面算法速度较快,可以满足计算机视觉系统的要求。
The haze weather causes the image obtained by the imaging device to degrade,which affects human visual perception and computer vision feature extraction.Computer vision system has high requirements on the speed of defogging,so a fast image defogging algorithm is proposed.Firstly,the airlight is estimated by Gaussian blur based on the foggy image degradation model,the atmospheric light value is estimated based on the overall image brightness,and finally the defogged image is recovered based on the foggy image degradation model.The experimental results show that in the aspect of subjective visual effects,the defogging effect is obvious and the image quality is significantly improved.In the aspect of defogging speed,the algorithm in this paper is fast and can meet the requirements of computer vision systems.
作者
崔建伟
王冬青
刘金燕
CUI Jianwei;WANG Dongqing;LIU Jinyan(School of Automation,Qingdao Shandong 266071,China;Electrical Engineering,Qingdao Shandong 266071,China)
出处
《自动化与仪器仪表》
2021年第1期9-11,16,共4页
Automation & Instrumentation
基金
国家自然科学基金资助项目:复杂网络拓扑与参数的辨识(No.61573295)
国家自然科学基金资助项目:基于数据特征的多模态过程辨识建模方法(No.61873138)。
关键词
图像去雾
高斯模糊
环境光
大气光值
image defogging
gaussian blur
airlight
atmospheric light
