P‐5.6: Synergistic Low‐Light Image Enhancement: A Fusion of Dark Channel Dehazing and K‐means Clustering

Abstract

In critical areas such as security surveillance and nocturnal navigation, low‐light image enhancement is crucial. However, current algorithms face challenges such as uneven brightness, noise amplification, and haze artifacts, which compromise image quality and utility. This paper introduces a novel enhancement algorithm that integrates dark channel prior with K‐means clustering, resulting in uniform brightness and reduced noise. A dark channel dehazing technique has been enhanced to further reduce haze artifacts. Our algorithm's superiority in restoring true colors and outperforming existing algorithms has been demonstrated through comprehensive evaluation on the DICM dataset and low‐light color charts, affirming its effectiveness in low‐light scenarios.