Image Haze Removal Algorithm Using a Logarithmic Guide Filtering and Multi-Channel Prior

In reality, the quality of an image is generally affected by haze. To obtain a well-quality image, removing haze is a hot issue on theory and application. This paper proposes a new algorithm to remove haze of hazy images. In the algorithm, first, the ambient illumination is estimated by a logarithmic guide filtering that can reserve the characteristics of the bright source areas and improve the dark source areas of the hazy image. Second, to overcome the defect of dark channel prior (DCP) and the over-brightness of the bright channel prior (BCP), two models with two parameters are introduced to improve the DCP and BCP, called multi-channel prior method. At the same time, a self-adaptive method is presented to compute the values of the two parameters. At last, based on the multi-channel prior, a self-adaptive method is proposed to compute the transmission mapping value. Further, four classes hazy images are employed to test the proposed method. The experimental results carried out on the public databases demonstrate that the proposed algorithm can outperform the current state-of-the-arts, including more effective defogging, clearer visibility and richer details.

Abstract licence: CC BY-NC-ND

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.