Image Dehazing Algorithm Based on Dark Channel Sky Region Segmentation

In response to the distortion issues encountered by the dark channel prior algorithm when processing white regions such as the sky, as well as the challenges of low computational efficiency and loss of edge details, this paper proposes an improved adaptive sky region segmentation algorithm for high-definition image dehazing. This algorithm accurately distinguishes between sky and non-sky areas in the image using adaptive sky region segmentation techniques. It then precisely estimates the transmission rates of the non-sky regions based on image gradient information and compensates for the transmission rates in the sky regions. Furthermore, the transmission rates are refined through fast guided filtering. Experimental results demonstrate that this algorithm effectively enhances the dehazing effect while preserving details, and the restored images significantly mitigate halo artifacts and distortion in bright areas.