一种边缘优化的暗通道去雾算法

针对暗通道先验去雾算法大部分的时间都消耗在对透射率的优化上的问题, 对暗通道去雾算法进行了改进, 提出了边缘优化的暗通道去雾算法。使用边缘算子从粗略估计的透射率中提取边缘, 对边缘及周围扩展区域内的像素采用差值抠图法优化图像中场景深度变化明显的区域的透射率。实验结果证明, 边缘优化的暗通道去雾算法在得到与原始算法基本一致的去雾结果的同时, 平均计算时间仅为原算法时间的60%左右, 有效减少了计算量, 提高了去雾算法的运算速度。

Dark channel prior dehazing algorithm based on edge optimization

Most of the computation time was consumed in the step of optimization the transmittance in the dark channel prior dehazing. Aiming at this problem, this paper modified the dark channel prior dehazing algorithm, and proposed a new dark channel prior algorithm based on edge optimization. The new algorithm firstly extracted the edges from crude estimated transmittance image. Then, for the pixels on and near the edges, the difference matting optimization algorithm was performed on the regions that the depth changed significantly. The experimental results demonstrate that the performance of the proposed algorithm is substantially the same as the original one. The average computation time of the new algorithm is around 60% of the original one. Therefore, it can reduce the computation load effectively and improve the speed of the dehazing algorithm.