IMAGE DEHAZING USING DEEP LEARNING ALGORITHMS

Abstract

Image dehazing is a crucial task in computer vision that aims to restore clear images from hazy or foggy conditions. Atmospheric particles such as dust, smoke, and water droplets scatter light, leading to reduced visibility, low contrast, and color distortion in captured images. This degradation negatively impacts various real-world applications such as autonomous driving, surveillance systems, remote sensing, and medical imaging. Traditional image dehazing methods rely on handcrafted priors like the Dark Channel Prior (DCP), which often fail in complex scenes and require manual tuning. With the advancement of deep learning, data-driven approaches have emerged as powerful alternatives for image dehazing. Convolutional Neural Networks (CNNs) and Generative Adversarial Networks (GANs) can learn complex mappings between hazy and clear images, enabling more accurate restoration. This paper presents a deep learning-based image dehazing framework that utilizes convolutional layers to extract features and reconstruct haze-free images. The model is trained on large-scale datasets containing paired hazy and clear images.The proposed system improves visibility, enhances contrast, and preserves color fidelity. Experimental results demonstrate that deep learning-based methods outperform traditional approaches in terms of Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM). This study highlights the effectiveness of deep learning techniques in addressing real-world dehazing challenges and improving visual quality in degraded images.