Multi-Objective Adaptable Atmospheric Light and Depth Map Quantization Level Dehazing With CSA
8 Pages Posted: 6 Jan 2020
Date Written: December 30, 2019
Abstract
Outdoor natural images captured in turbid weather are subject to produce low visibility which may cause fatal problem in computer vision applications. Single image visibility improvement is the most difficult of all model used in visibility improvement and mostly depends on restoration based optical image formation model. Existing dehazing techniques estimate atmospheric light (AL) and transmission by some prior information. In this work, we examined the strength of the Cuckoo Search Algorithm(CSA) in tuning a new multi-objective image performance function PCS for adaptable dehazing which estimates and balance between noise, contrast and geometrical information of image . Two parameters, AL and Otsu’s quantisation thresholding level in Depth Map (DM), are optimised with levy steps in search space of CSA to produce best dehazed image . GT O-Haze, DerainNet, Frida synthetic dataset have been used for evaluations and our method is compared with ten state-of-the art methods qualitatively and qualitatively for dehazing obtaining satisfactory results.
Keywords: image recovery model, Quality Assessment, CSA, Dehazing, extinction coefficient, fitness function, PCS, levy flight
Suggested Citation: Suggested Citation