结合WLS滤波与还原控制因子的图像去雾①王伟鹏1,项文杰2,刘新芳11(闽南科技学院光电信息学院,泉州362332)2(厦门大学海洋与地球学院,厦门361102)通信作者:项文杰,E-mail:yyxiangwj@163.com摘要:雾天退化图像的复原过程中,针对大气光幕和大气亮度估计不准确导致光晕效应、偏色现象和对比度不足等问题,提出一种结合WLS(weightedleastsquare)滤波与还原控制因子的去雾算法.首先分析WLS滤波器的原理和性能,并用于大气光幕的有效提取;其次利用Sobel算子检测二值化图像边缘,将边缘数目与像素均值同时作为四叉树空间索引的依据,提高大气亮度的估计准确性;最后分析天空出现颜色失衡现象的原因,引入还原控制因子改善视觉效果.实验结果表明,去雾后图像的平均梯度整体提高58.03%,信息熵提高2.88%,运行时间节省50%以上.该方法对含有浓雾、薄雾以及天空等深度复杂的远景图像、近景图像均能得到高对比度、可视度和色彩保真度的恢复效果.关键词:图像去雾;加权最小二乘法滤波(WLS滤波);图像复原;图像增强;大气散射模型引用格式:王伟鹏,项文杰,刘新芳.结合WLS滤波与还原控制因子的图像去雾.计算机系统应用,2023,32(2):303–309.http://www.c-s-a.org.cn/1003-3254/8947.htmlImageDehazingBasedonWLSFilteringandRestorativeControllingFactorWANGWei-Peng1,XIANGWen-Jie2,LIUXin-Fang11(CollegeofOptoelectronicInformation,MinnanScienceandTechnologyUniversity,Quanzhou362332,China)2(CollegeofOceanandEarthSciences,XiamenUniversity,Xiamen361102,China)Abstract:Astheinaccurateestimationofatmosphericlightcurtainandatmosphericlightresultinthehaloeffect,colordistortion,andlowcontrastintheprocessofhazeimagerestoration,adehazingalgorithmbasedonweightedleastsquare(WLS)filteringandrestorativecontrollingfactorisproposed.Firstly,thisstudyanalyzestheprincipleandperformanceoftheWLSfilter,whichcanbeutilizedtoeffectivelyestimatetheatmosphericlightcurtain.Secondly,withtheassistanceoftheSobeloperator,thebinaryimageedgesaredetected.Thenumberofedgesandthemeanvalueofpixelsaretakenasthebasesofthequad-treespaceindex,whichimprovestheestimationaccuracyoftheatmosphericlight.Finally,accordingtothecausesofcolordistortionintheskyarea,arestorativecontrollingfactorisintroducedtoimprovevisualeffects.Experimentalresul...
