Improved Haze Removal Method using Proportionate Fusion of Color Attenuation Prior and Edge Preserving

摘要

Nowadays, fog and haze are becoming a global challenge. Images captured under the hazy condition have poor contrast and corrupted colour. Such images limit the visibility and thus hinder the way for the computer vision purposes like video observation, entity recognition. Such hazy images affect the usual working in the transportation sector e.g. trains, ships etc. One of the prominent factors affecting the outdoor vision applications is poor vision and thus, it may result in an intangible loss like safety. The image captured shows this behaviour because of the air light. This paper analyses the widely used techniques for haze removal from images alias color attenuation prior based haze removal and haze removal with edge preserving. Also, the paper proposes the better haze removing techniques using fusion based approach which gives better quality haze removal compared to the bench mark haze removal techniques. Experimental results tested with NIQE have proved the worth of proposed methods.