Video Dehazing Based on DCP and Type-2 Fuzzy Sets with a Guided Filter

Abstract

Video dehasing refers to techniques for improving the visibility and quality of videos recorded under unfavourable weather conditions like haze or fog. This research presents a novel approach to improve the quality and exposure of these kinds of films. Using a regulated filter, type 2 fuzzy sets, and the dark-channel-prior, the method creates videos with higher resolution. The dark-channel prior estimates atmospheric light conditions and also transmission maps for each frame in the video. By integrating type-2 fuzzy sets, the approach effectively handles uncertainties and blurring in out-of-focus scene information, improving the robustness and accuracy of the dehazing process. Additionally, a guided filter refines the transmission maps and enhances the video images, reducing artefacts and improving visual quality. To gauge the effectiveness of the approach, quality evaluation metrics like Peak-Signal to Noise-Ratio (PSNR), Structural-Similarity Index (SSIM), Lightness-Order-Error (LOE), and Naturalness-Image-Quality-Evaluator (NIQE) are utilized. This research contributes to the further development of image equalization techniques and their practical applications across various fields.