Image Dehazing Based on Luminance Stretching
Geet Sahu, Ayan Seal · 2019
In this work, a method is proposed to restore the visual effects of hazy images with sky using dark channel prior (DCP) and luminance stretching (LS). The transmissions for the non-sky and sky regions of a hazy image is calculated by DCP and LS respectively. The transmission of the hazy image is computed by combining the transmissions obtained using DCP and LS based on soft segmentation. The proposed method along with two state-of-the-art methods is evaluated on 125 randomly selected images from BSDS500 database. The qualitative and quantitative results depict that the proposed method outperforms two state-of-the-art methods.