Dehazing of Satellite Images using Adaptive Black Widow Optimization-based framework

• DOI: 10.1080/01431161.2021.1910367
• Publication Date: 2021-04-06T11:06:19Z
• Authors: Shilpa Suresh; Ragesh Rajan M.; Jagalingam Pushparaj; Asha C; Shyam Lal; Chintala Sudhakar...
• Keywords: Adaptive Black Widow Optimization-b.; Adaptive Black Widow Optimization; Information Systems not elsewhere classified; de-hazing satellite imagery; Contrast Enhancement; 02 engineering and technology; Weighted Variance Guided Filter; filter-based image refinement scheme; Environmental science; Mathematical Sciences not elsewhere classified; Multispectral and Hyperspectral Image Fusion; Engineering; Space Science; Environmental Sciences not elsewhere classified; fusion-based transparency function; Dehazing; Media Technology; 0202 electrical engineering, electronic engineering, information engineering; image quality; Computer Graphics and Visualization Techniques; Haze Removal; Image Fusion; Image Enhancement Techniques; 006; 600; Geology; FOS: Earth and related environmental sciences; Remote sensing; DCP; Computer Graphics and Computer-Aided Design; Computer science; two-stage image decomposition; ABWO; Aerospace engineering; Satellite; Computer Science; Physical Sciences; method; Medicine; Computer Vision and Pattern Recognition; WVGF; Biological Sciences not elsewhere classified; Single Image Restoration

Haze is a common atmospheric disturbance that adversely affects the quality of optical data, thus often restricting their usability. Since these effects are inherent in process spaceborne Earth sensing, it important to develop effective methods remove them. This work proposes novel method for de-hazing satellite imagery and outdoor camera images. It developed by modifying transmission map used Dark Channel Prior (DCP) method. A Weighted Variance Guided Filter (WVGF) introduced enhancing image quality, which included two-stage decomposition fusion process. The also optimally combines radiance components along with an additional stage modelling fusion-based transparency function. final guided filter-based refinement scheme incorporated improve processed quality. optimal tuning image-dependent parameters at various stages achieved using newly proposed Adaptive Black Widow Optimization (ABWO) algorithm, makes fully automatic. Qualitative quantitative performance analyses, results compared other state-of-the-art methods. experimental reveal performs better as others, independent haze density, without losing natural look scene.