退化图像的去雾方法  

Defogging method for degraded image

作　　者：黄黎红[1] 

机构地区：[1]莆田学院电子信息工程系,福建莆田351100

出　　处：《红外与激光工程》2010年第5期985-988,共4页Infrared and Laser Engineering

基　　金：福建省自然科学基金计划项目(S0750014)

摘　　要：由于大气的散射作用,雾天的大气退化图像具有对比度低、景物不清晰的特点,给交通系统及户外视觉系统的应用带来严重的影响。因此,对雾天低对比度图像进行去雾处理有着重要的意义。在大气物理退化模型的基础上,根据雾对景物退化与深度的指数关系,提出了一种去雾算法。该方法仅需单幅退化图像,采用一种景物深度估计法估算景物深度,并结合运用了亮度调整,最后利用亮度增益曲面对原彩色图像RGB三颜色分量同比增强,保证了色调恒定不失真。实验结果表明:该算法去雾效果显著,操作简单,且色彩自然,速度快,可以实现实时视频去雾的要求。Atmospheric degraded images captured in fog weather were of poor contrast and low visibility as a result of atmosphere scattering.These images brought serious effects in the application of traffic system and outdoor visions system.Therefore,defogging for images of poor contrast which were captured in fog weather is of significance.Based on atmospheric physical degrading model,a novel defogging method was proposed according to the exponential relationship between the degradation and depths of the scene points.The single degraded image was needed only.And a new depth estimation method was used to estimate scene depth.Combined with adjustments on illumination,the three RGB components of original images were enhanced by illumination gain surface,which ensured the hue be constant and undistorted.Experimental results show that the new method can defog images effectively,and it possesses some characteristics such as simple operation,natural color and quick process.Furthermore,the defogging real-time videos can be realized.

关 键 词：去雾 物理模型 图像复原 景深 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]