Image dehazing using dark channel prior-based adaptive filtering and contrast-limited adaptive histogram equalization

Abstract

Image dehazing is essential for vision enhancement, particularly in challenging environments such as mountainous and fog-prone regions. However, existing dehazing methods often suffer from high computational complexity, inaccurate transmission estimation, colour distortion, edge degradation, and noise amplification. While learning-based approaches and dark channel prior (DCP) techniques have demonstrated promising performance, they still struggle to balance visual quality, robustness, and efficiency. This paper proposes a novel Dark Channel Guided Adaptive Histogram Equalization (DCGAHE) method for single-image dehazing. The proposed approach jointly leverages DCP for haze removal, adaptive guided filtering for transmission map refinement and edge preservation, and contrast-limited adaptive histogram equalization (CLAHE) for controlled contrast enhancement. By integrating guided filtering with CLAHE, DCGAHE enhances image clarity and contrast while effectively suppressing noise amplification and preserving structural details. The performance of DCGAHE is evaluated using five quantitative metrics as used in literature for better comparison. Experimental results demonstrate that DCGAHE consistently outperforms state-of-the-art dehazing techniques in terms of visual quality, edge fidelity, and contrast enhancement, producing more natural and visually pleasing dehazed images.