Real-time defog model based on visible and near-infrared information

It is important to guarantee that images circulating in mobile cloud of smart city are not obscured by fog. Based on near-infrared's ability to penetrate fog, this paper puts forward a 2-step real-time defog model for camera: putting forward the infrared - blue light intensity difference factor...

Full description

Saved in:
Bibliographic Details
Published in2016 IEEE International Conference on Multimedia & Expo Workshops (ICMEW) pp. 1 - 6
Main Authors Zhang Jingyun, Ding Yifan, Yang Yi, Sun Jiasong
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.07.2016
Subjects
Guided image filtering
Image dehazing
Image fusion
Indexes
Near-infrared
Optical imaging
Optimization
Real-time systems
Telecommunications
Online AccessGet full text

Cover

More Information
Summary:It is important to guarantee that images circulating in mobile cloud of smart city are not obscured by fog. Based on near-infrared's ability to penetrate fog, this paper puts forward a 2-step real-time defog model for camera: putting forward the infrared - blue light intensity difference factor and using dark channel prior to estimate haze distribution map; fusing near-infrared and visible information based on the haze distribution and then, adopting downsampling, fast fuzz algorithm and fast guided filter to remove the artificial effect and improve the efficiency. This simple and efficient method is rid of complicated physical models and cumbersome manual operations while provides results with richer color, sharper details than other algorithms, achieving accuracy, real-time, fidelity and automation at the same time. Camera based on this model is an excellent multimedia application for mobile cloud of smart city.
DOI:10.1109/ICMEW.2016.7574749