公路场景去雾技术研究  

Study of defog technology for road scene

作　　者：刘红军 宫爱玲[1] 周燕[2] 王新伟[2] 

机构地区：[1]昆明理工大学,云南昆明650093 [2]中国科学院半导体研究所光电系统实验室,北京100083

出　　处：《光学技术》2010年第4期554-559,共6页Optical Technique

基　　金：863计划(2007AA11Z238)资助项目

摘　　要：提出了一种应用单目机器视觉技术并结合Hough变换消失线检测和边界检测技术来构建距离深度图的方法,并应用衰减-空气光去雾模型对图像进行了恢复。结果表明,采用该方法恢复的道路图像色彩自然,模型与实际道路完全匹配,对于不同距离的景物都具有良好的恢复效果。In order to remove weather effects from images,monocular camera machine vision,Hough transform and edge detection are combined to construct the scene depth map. It is assumed that the brightness of each pixel at the receiver is composed of scene point attenuation irradiance and airlight illumination scattering intensity,by removing airlight illumination scattering and removing particles to restore the image degraded in foggy day,the color of image restored by this method is natural,the model and real road are matched accurately,the different distance scene has satisfied effect.

关 键 词：摄像机标定 消失线 衰减-空气光模型 单目视觉 HOUGH变换 边界检测 去雾 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]