改进暗通道先验的雾霾天图像复原算法

为解决暗通道先验在雾霾天图像复原过程中存在的不足,提出了一种改进的快速算法,即采用自适应邻域求取原图像的暗通道,解决了固定邻域在局部区域的错误估计问题;结合极大值滤波与双边滤波计算透射率,有效降低了运算量和运行时间;最后利用区域像素的平均值替代单个最大值,获取更加准确的空气光亮度,从而使改进算法对雾霾去除的视觉效果更加...

Image Restoration Algorithm in Haze Days Based on Improved Dark Channel Prior

In order to solve the deficiency of dark channel prior in the process of image restoration in haze days, an improved fast algorithm was proposed, that is, an adaptive neighborhood was used to obtain the dark channel of the original image, which solved the problem of error estimation of fixed neighborhood in local area. Combining the maximum filtering and bilateral filtering to calculate the transmittance, the amount of calculation and running time are effectively reduced. Finally, the average value of regional pixels is used to replace the single maximum value to obtain more accurate air brightness, so that the visual effect of the improved algorithm for haze removal is more outstanding, and the execution time of the algorithm is greatly reduced.