Haze Removal using Computer Vision Technique
On the basis of the dark channel prior (DChP) and guided filter an efficient technique of haze removal proposes in this paper for the single image. The statistics of an open-air haze free picture are the DChP. Depending on keen observation, in many local patches, the intensity of few pixels of open-air haze free pictures is very less in at-least single shading channel. With the help of DChP, the heaviness of haze can be estimated along with the recovery of the high-quality dehaze image. Results on different hazy images reflect the strength of the proposed algorithm. As a byproduct of haze elimination, a depth map having good-quality can be obtained. In addition, the guided filter uses for transmission refinement helps in reduction of time complexity.
Keywords - Haze Removal, Dark Channel Prior, Atmospheric Light, Scene Radiance.