<p>One of the most significant challenges in image processing is the task of restoring degraded images, due to varying weather conditions such as the adverse effects of atmospheric factors on outdoor images’ visibility. Haze introduces complexities in diverse computer vision applications, emphasizing the importance of contrast enhancement and visibility restoration in such foggy images. This paper conducts an in-depth survey of dehazing techniques utilizing convolutional neural networks (CNNs). Image dehazing, which enhances visibility and contrast in hazy images, has garnered substantial attention in the last years. The survey commences by comprehensively discussing the physical models, datasets, challenges, popular categories of dehazing methods, and prevalent evaluation metrics used in dehazing research. These components serve as foundational elements for understanding the underlying principles and evaluating dehazing algorithms. The study subsequently offers an analytical exploration of various deep learning techniques applied to image dehazing, aiming at providing an intuitive comprehension of the pivotal techniques for haze removal. Through quantitative and qualitative experiments against diverse baseline methods, the paper establishes performance benchmarks and allows for the assessment of different approaches. Finally, the paper identifies unresolved challenges, presenting a roadmap for future research and improvements in dehazing techniques.</p>