ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

Integrated Approach for Solving Haziness in an Image Using Dark and Bright Channel Prior

Journal: International Journal of Science and Research (IJSR) (Vol.5, No. 5)

Publication Date:

Authors : ; ;

Page : 143-147

Keywords : Dark channel prior; Bright Channel Prior; Image Dehazing; Soft Matting; Atmospheric Light; Haze Model; Fast Matting; Image Enhancement;

Source : Downloadexternal Find it from : Google Scholarexternal

Abstract

Haze is a natural phenomenon that reduces visibility, clouds scenes, and changes colour. Since image quality degrades it is an irritating issue for photographers. The removal of haze, called dehazing, Haze is also a threat to the reliability of many applications, like outside surveillance, object discovery, and aerial imaging. Removing haze from pictures is important in computer vision/graphics. But haze removal is tremendously challenging due to its ambiguity that is mathematical. Here we propose a new prior called Bright Channel Prior to de -haze single picture combining with the Dark channel prior. The bright channel prior, which inspired by the dark channel earlier, is a statistic of haze-free images which can be outside. It really is centered on a key observation - local patches in haze-free pictures that were outside include some pixels which have very low intensity in the absolute minimum of one colour channel. Using these priors with the imaging model that is haze, we can estimate the depth of the haze and regain a high quality haze-free image. Results on many different outdoor haze images present the power of the earlier that is projected. On the other hand, the DCP scheme has time consuming trouble as a result of matting that is soft. By using fast matting system using large kernel matting Laplacian matrices instead of time consuming soft matting in this paper the goal of planned dehazing procedure can also be able to improve quality and reduce calculation time of the picture simultaneously.

Last modified: 2021-07-01 14:37:34