In the realm of image processing, haze remains a pervasive challenge that significantly degrades the visibility and clarity of outdoor images. Such degradation not only hinders human perception but also impairs the performance of various computer vision applications. Existing de-hazing techniques often struggle with the complex interplay of light and colour, leading to suboptimal restoration of true scene radiance. This work addresses the limitations of current methods, which frequently rely on insufficiently robust priors, neglecting the nuanced variations of atmospheric light scattering. These challenges are compounded when images lack ground truth data, rendering the assessment of enhancement quality particularly difficult. Consequently, there is a pressing need for an approach that adapts to varying haze densities and provides a reliable quality metric in the absence of reference images. Responding to this, the presented research proposes a novel framework for single-image de-hazing that introduces improved pre-processing steps including adaptive equalization, neighbourhood sharpening, and gamma correction within the LAB colour space model. A multi-colour space-based model is also introduced, utilizing Light Channel Prior and Dark Channel Priors to estimate global atmospheric contrast more accurately. The methodology is comprehensive, employing quantitative metrics such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), processing time, and the fog-aware density evaluator (FADE) score for performance evaluation.