An single image dehazing algorithm using sky detection and segmentation

Image dehazing has always been a focus of wide attention in recent years with great breakthrough. However, the existed algorithms may produce noise and color distortion in the process of image restoration, especially in the sky region. In this paper, an improved dark channel prior algorithm is proposed, which detects the sky firstly and divides the image into sky region and nonsky region, and then estimates the transmissions of the two parts separately, followed by the combination with a refining step. The results show that more natural images can be restored in the process of image dehazing, whether there is a sky area or not.