ENHANCEMENT OF FAR-REGION HAZY IMAGE USING MULTIEXPOSURE IMAGE FUSION

Abstract

Poor climate conditions can decrease visibility of images obtained outside, reducing their visual quality. The image processing task concerned about the alleviation of this impact is known as image dehazing. Many of such dehazing techniques have been introduced. However these methods cannot dehaze the far haze and high hazy images. In this work a new algorithm is designed to overcome these drawbacks. In this method, first the original hazy image is under-exposed through the sequence of gamma-correction activities. Then the resulting images with different exposure can be fused into a haze free image through a multi-scale Laplacian pyramid blending method. Experimental results on a set of foggy images determine that the proposed method better preserves the visibility of faraway hazy images and images that are highly affected by haze.