An Effective and Efficient Dehazing Method of Single Input Image

The quality of an image may be degraded seriously when it is captured in a foggy weather condition. In this paper, an effective and efficient dehazing method is proposed for a single input image by combining the dark channel prior information and a low-light image enhancement model. First, the dark channel is derived via two minimum operations. After estimating the atmospheric light, the transmission is initialized according to the property of aerial perspective. In terms of the atmospheric light, a bound constraint is computed further to refine the transmission. Finally, a high-quality image is obtained via the haze image model. Experimental results demonstrate the effectiveness and efficiency of the proposed method.