Enhanced Underwater Image Dehazing Using an Improved Dark Channel Prior Method

Underwater image dehazing is essential to counteract quality issues from absorption and scattering. In this paper, we propose an improved dehazing technique based on the Dark Channel Prior (DCP) method, called the Improved Dark Channel Prior (IDCP). The IDCP combines a novel airlight estimation strategy, termed Maximum Ratio Estimation (MRE), and an enhanced transmission map estimation method, termed Red Channel Retaining (RCR). Unlike traditional methods, IDCP incorporates the red channel, often overlooked in underwater imaging due to rapid attenuation, to improve color restoration and image clarity. We evaluated the proposed method using the Underwater Image Enhancement Benchmark (UIEB) dataset. Experimental results demonstrate that IDCP significantly enhances the color correction, contrast, and overall image quality, compared to conventional DCP technique. Additionally, quantitative metrics such as entropy and mean squared error indicate a substantial improvement in the restored images, highlighting the effectiveness of our approach in real-world underwater imaging scenarios.