A Comparative Study of Image Dehazing Algorithms

Haze poses a great challenge in modern-day applications. Removing haze is a challenging task because haze varies with the depth of the scenes in the image. Many automated systems like surveillance systems, object tracking, etc. use the dehazing methods internally to improve their overall performance in a hazy environment. Hence, recovering the original image from the hazy image is a crucial task. This paper provides a detailed survey of many algorithms that have been proposed for haze removal, which refines the image, either by using statistical observation of the scene or by network-based learning method. These algorithms have been assessed quantitatively and visually to compare based on their dehazing potential. The paper presents a comparative study of the various image dehazing methods having a different way of estimation of transmission and atmospheric light. This paper reviews different prior-based technique and learning-based technique image dehazing algorithms.