Font Size: a A A

Research And Optimization Of Image Dehazing Algorithm Based On Atmospheric Scattering Model

Posted on:2021-11-21Degree:MasterType:Thesis
Country:ChinaCandidate:Y FuFull Text:PDF
GTID:2518306200953129Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
Nowadays,fog and haze weather occur frequently in more and more areas,the propagation of light under such atmospheric conditions is influenced by the absorption and scattering of dielectric particles,as the energy decays,the direction of the light path will be changed to some extent.The images obtained by the corresponding electronic equipment are often blurred,with low contrast and color attenuation distortion.This degradation has seriously affected the normal work of various military and people's livelihood systems based on image data.Therefore,it is of great research significance and practical value to carry out effective defogging operations for this kind of degraded images to restore image quality.Firstly,according to whether the processing method is based on the physical model,this paper introduces the principles of the main research results related to image defogging at home and abroad,analyzes the advantages and disadvantages of their processing effects,and expounds the degradation principle and degradation performance of foggy images.Next,this paper focuses on the research of the widely used dark channel prior dehaze algorithm,from its theoretical content,the establishment conditions of the prior and the verification experiments,the ways to obtain the parameters.After a comprehensive analysis of the processing process of the algorithm,the shortcomings of the algorithm are summarized.When there are large bright areas and sudden changes in depth of field in foggy images,color deviation and halo effect are easy to appear in the results processed by traditional fogging methods.In order to solve these problems,an improved algorithm is proposed to set the range of transmittance estimation based on the boundary constraint and to modify the transmittance by setting the threshold.In the phase of atmospheric light estimation,the improved algorithm combines automatic and manual estimation,which provides a solution for users to further adjust the defogging results according to the requirements.For the estimation of transmittance,the upper and lower limits of the corresponding transmittance are obtained through the boundary constraint of the radiator to replace the preset fixed values in the traditional algorithm,so as to more accurately limit the value range of the transmittance.Then,the threshold value is calculated to determine whether the pixel is in the same depth of field range,and according to the intensity difference value ratio,the corresponding transmittance correction is made adaptively to optimize the parameter estimation.Through multiple sets of different scenarios,different data sources of fog images processing experiments comparison,using subjective and objective analysis show that the algorithm can achieve good image dehaze effect,the restore clear images,enhance the visual effect and availability at the same time,effectively avoid the diagram appears after the bright area in dealing with color migration artifacts and scenes depth mutation position problems of the halo effect.
Keywords/Search Tags:image dehazing, atmospheric light value, transmittance, boundary constraint, adaptive correction
PDF Full Text Request
Related items