基于灰度特征值分割的暗通道先验去雾增强算法  

Dark channel prior defogging enhancement algorithm based on gray scale eigenvalue segmentation

作　　者：周研 于萍[1] ZHOU Yan;YU Ping(School of Mathematics and Computer Science,Jilin Normal University,Siping 136000,Jilin,China)

机构地区：[1]吉林师范大学数学与计算机学院,吉林四平136000

出　　处：《智能计算机与应用》2024年第7期71-78,共8页Intelligent Computer and Applications

摘　　要：暗通道先验算法去雾的效果较好,但是去雾后的图像普遍存在光晕现象和亮度偏低等问题。基于双边滤波的多尺度Retinex算法去雾的效果一般,但是可以提高图像的亮度和清晰度,因此本文提出基于灰度特征值分割的暗通道先验去雾增强算法,首先基于灰度特征值分割天空区域来更好的求出大气光值;其次,改进透射率模型后用引导滤波对其进行细化;最后,用容差值对亮区域的透射率进行调整。去雾后的图像利用基于双边滤波的多尺度Retinex算法提高图像的亮度和清晰度,用sigmoid函数调整对比度,最后线性拉伸得到最后的增强图像。实验结果表明,基于灰度特征分割的暗通道先验去雾增强算法可使图像具有更好的视觉效果,图像的亮度、清晰度较好,细节保留也更加完整。The prior algorithm of dark channel has a better effect on fog removal,but there are some problems in the image after removing fog,such as halo phenomenon and low brightness.And the effect of multi-scale Retinex algorithm based on bilateral filtering is not good,but can improve the brightness and clarity of the image.Therefore,in this paper,we propose an enhanced algorithm of dark channel prior defogging based on gray-scale eigenvalue segmentation,firstly,the algorithm divides the sky region based on the gray level eigenvalue to get the atmospheric light value,then the transmittance model is improved,it is then refined by the guiding filter,finally,the transmittance of the bright region is adjusted by the tolerance value.The brightness and clarity of the defogged image are improved by using the multi-scale Retinex algorithm based on bilateral filtering,then adjust the contrast using the sigmoid function,finally,the final image is obtained by linear stretching.The experimental results show that,Gray-level feature segmentation based prior enhancement algorithm for Dark Channel defogging can make the image have better visual effect,the brightness and clarity of the image are better,the details are more complete.

关 键 词：暗通道 分割天空区域 改进透射率 增强图像 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]