高斯自适应多尺度加权滤波去雾算法  

GAUSSIAN ADAPTIVE MULTI-SCALE WEIGHTED FILTERING DEHAZING ALGORITHM

作　　者：杨洋 宋春花 Yang Yang;Song Chunhua(School of Information and Computer,Taiyuan University of Technology,Jinzhong 030600,Shanxi,China)

机构地区：[1]太原理工大学信息与计算机学院,山西晋中030600

出　　处：《计算机应用与软件》2022年第3期187-192,265,共7页Computer Applications and Software

基　　金：国家自然科学基金项目(51505324)。

摘　　要：为了达到良好的图像去雾效果,提出一种高斯自适应多尺度加权滤波去雾算法。通过多尺度最小值加权滤波得到暗通道图像,建立最小通道与高斯函数的关系,线性约束后并经过自适应参数对像素灰度值的调整得到粗级透射率,紧接着对得到的粗级透射率图像进行多尺度加权引导滤波得到优化透射率,结合加权大气光强并依据大气散射模型对图像进行去雾复原处理。实验结果表明,该方法有效地将单幅有雾图像进行了处理,与其他经典算法相比较得到的图像细节显示效果好,很好地恢复了场景的对比度,增加了图像的可见度,具备一定的优异性。In order to effectively dehaze the image,a Gaussian adaptive muti-scale weighted filtering dehazing algorithm is proposed.The dark channel image was obtained by muti-scale minimum weighted filtering.After that,the relationship between the minimum channel and Gaussian function was established.The coarse transmittance was gotten through the adjustment of pixel grey value by adaptive parameters after the linear constraint and the optimal transmittance could be obtained by muti-scale weighted guided filtering of the coarse level transmittance diagram.With the weighted atmospheric light intensity,the image was dehazed and recovered according to the atmospheric scattering model.The experimental results show that the method in this passage can process the single foggy image effectively.Compared with other classical algorithms,this method is able to achieve more effective detail display of the image,and can recover the contrast ratio well and increase the visibility of the images,which has certain excellence.

关 键 词：图像去雾 多尺度加权 高斯函数 自适应标准差 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]