基于AS模型和自适应双边滤波快速去雾算法

Fast Removal Fog Algorithm Based on the Adaptive Bilateral Filtering

为了获得雾天图像原始信息,减少颜色失真及边缘缺失等现象发生,提出了一种基于AS模型和自适应双边滤波快速去雾算法。由双边滤的波平滑特性推出大气耗散函数,并结合估测的大气光强实现了雾天退化处理;针对雾天图像明亮失真区域,引入了辅助函数对透射率进行自适应调控及其优化,实现了对雾天图像颜色失真和局部弱化的处理。实验仿真结果表明,文中算法处理的4组不同景深下的雾天图像整体清晰度要有优于其他算法,运行时间减少,最快可节省54.9480s。在细节处理上更细腻,边缘信息保留更加完整,尤其是在景深突变的边缘处无Halo效应,颜色失真得到了较大改善。

In order to obtain the original information of the fog image and reduce the color distortion and marginal loss, a fast removal algorithm based on AS model and adaptive bilateral filtering algorithm is proposed. First, the atmospheric dissipation function is estimated by the bilateral filtering smoothing characteristics, and the fog image removal processing is realized by combining with the estimation of atmospheric light intensity value. Then, concerning the distortion of the bright portion, the color distortion and local weakening are further realized by introducing the auxiliary function to the adaptive control and optimization of the transmittance. The simulation results show that the 4 groups of different depth under the fog images in the new algorithm, the overall definition is superior to other methods, computing time occupies obvious advantages, and the fastest can save 54. 9480 seconds. The details of the treatment are more delicate, edge information retains more complete, especially at the edge of the depth of field mutation without halo effect, the overall color distortion improves a lot.