Chapter 20 - An efficient single image dehazing algorithm based on patch-wise transmission map estimation using Whale Optimization Algorithm

Abstract

In automated driving and video surveillance, image dehazing is a regular post-processing step, which can improve image visual quality that has been affected due to scattering and absorption of propagated light under hazy weather condition. To overcome this situation, we proposed single image dehazing method by using Whale Optimization Algorithm, called WOA-Dehaze. The proposed technique has three components: DHWT - Discrete Haar Wavelet Transform to Partition the supplied hazy image into sub-bands, estimating local atmospheric light, estimating each patch's transmission map, and fine-tuning with the Whale Optimization Algorithm. The information loss term, the image contrast term, and the fog density term are used to figure out the fitness cost for a single image. By reducing the cost function, the WOA-dehaze approach optimizes the gradient, image contrast, and information preservation. Extensive research on images of different scenes shows that the proposed WOA-Dehaze method outperforms existing fog removal techniques in terms of both quantitative accuracy and qualitative visual effect.