Haze Removal in Remote Sensing Images for Improved Data Analysis and Extraction
Satwinder Kaur, Bhawna Goyal, Ayush Dogra · 2022
Remote sensing imageries are extensively utilized for several applications such as categorization of crops, oil spill observation, cultivation of land, and forest fires. However, despite its crucial information content, the visibility of these crucial imageries gets corrupted due to attenuation or smattering because sometimes due to hazy climatic conditions atmospheric elements cause malfunctioning of scenic lights imageries. Therefore, appropriate dehazing schemes are mandatory for clearer imageries. In this paper, the technique that has been utilized is a compound fusion strategy that considers two input hazy imageries and dehazed the imagery by first intensifying the whiteness and contrast of it. For deriving the complete data and preservation of regions for proper perceptibility, significant features have been filtered by figuring three processes of weight maps which include first is for magnifying the luminance, second is for amplifying the quality of color and third is saliency which sharps the special features such as pixels as well as resolution. To reduce the artifacts which arise from the weight maps, the Laplacian pyramid methodology is being utilized. It also discusses some parameters that are responsible for the formation of the haze such as scenic depth, transmission map and atmospheric scattering model. Experimental results have also been revealed that the visibility of the recovered dehazed imageries is considerably improved by computing all steps which are shown in the figures.