A Hybrid Optimum Method for Removal of Haze using Weight Map Estimation and Fusion Strategy

The visibility of a scene is attenuated by external factors such as fog or haze, dimness and smog. It is also degraded by light scattering caused by minute dust fragments which absorb the light refracted in the atmosphere and the light emitted from an object. Hence this phenomenon clearly becomes an issue for image processing, as they are built for a normally clear field-view. Surveillance systems and outdoor object detection systems require the input image to be clear. Thus, the enhancement of visibility is very important for such applications. We propose a method that removes the haze in the input image in step by step process that includes contrast enhancement, white balancing, weight map estimation and multi-scale fusion techniques.