Dark Channel Prior based Image Dehazing with Contrast Enhancement

The visibility of images and videos captured by a camera is affected due to the haze particles in the atmosphere. The haze particles cause absorption and scattering of light which leads to poor image quality in outdoor images. The process of removal of haze from the images or videos is called dehazing, which is typically performed under the physical degradation model. In this work, image dehazing is performed by estimating two parameters namely transmission map and atmospheric light. The proposed work combines a histogram equalization approach with the technique of dehazing called dark channel prior. The results of the proposed method is validated using parameters SSIM, FADE and PSNR on a dataset called D-Hazy dataset and is found to be better when compared with other existing techniques.