Optimizing Visibility in Hazed Images: A Hybrid Approach with Dark Channel Prior, Guided Filtering, and Wavelets

Traditional dehazing methods often struggle to maintain color fidelity and detail preservation in dense haze conditions. This paper presents an advanced method for image dehazing, which integrates dark channel prior, guided filtering, and dual-tree complex wavelet transform to enhance visibility and image quality. The proposed approach begins with an initial stage of guided filtering to smooth the image while preserving edges, followed by the application of a dual-tree complex wavelet transform to decompose the image into frequency sub-bands. In these bands, a refined dark channel prior is applied to the wavelet approximation coefficients, enhancing the image by reducing the haze more selectively. The performance of the proposed method is quantitatively and qualitatively evaluated using metrics such as contrast, entropy, and perceptual fog density. The experimental results indicate significant improvements in image clarity and detail, making this method suitable for real-world applications where accurate image restoration is critical.