Single Fog Image Dehazing via Truncated Total Variation Method

Abstract

Existing dehazing methods are usually to appear visual problems. In the paper, we put forward a truncated total variation method (TTV) to eliminate haze. A histogram analysis is firstly developed to obtain global atmospheric light. Then, using an adaptive boundary constraint TTV to optimize the transmission properly. Finally, a new DCP is presented to remove haze. Shown in experimental results, our method can outperform existent methods on the visual effect.