Research On The Clearness Algorithm Of Sea Fog Images

Sea areas in China are one of foggy areas of the Pacific Ocean,and foggy days are more than 60% of the whole year.When it is foggy,sea visibility descends and the quality of the image collected by the maritime visual surveillance system degrades seriously,which directly affect the analysis and judgment of the image content,increase the difficulty of maritime safety monitoring,and reduce the safety of ships sailing in the fog.Therefore,it is of great practical significance to do intensive research into the clearness algorithm of sea fog images in order to enhance the reliability of maritime safety monitoring and the navigation safety of ships in the fog.Sea fog images often present the problems of low visibility,low contrast,color distortion,fuzzy edges and details,therefore,beginning with image enhancement and image restoration,a deep research on the clearness algorithm of sea fog images is made in this paper.The main research work includes the following aspects:Firstly,the principle and experimental results of histogram equalization algorithm and Retinex algorithm,which are commonly used in image enhancement defogging algorithm,are studied and analyzed.And their application effect in sea fog image clarification processing is tested.Aiming at these problems of color distortion,excessive and uneven enhancement,and halo phenomenon in the application effect,a sea fog image clarification algorithm based on improved Retinex theory is proposed.In the algorithm,the illumination component is estimated with improved guided filter in the HSV space,then,the reflection component is obtained and subjected to post-processing such as gray stretch,color recovery,and saturation compensation.The operations effectively enhance the image details,improve the color distortion,excessive enhancement and halo phenomenon,and make the overall quality of the image be improved to a certain extent.Secondly,aiming at this problem that the fog cannot be removed in nature and the enhancement effect is limited by using the defogging algorithm based on image enhancement,the clearness algorithm of sea fog images based on image restoration are studied further.Based on the degradation model of foggy images,a clearness algorithm of single foggy images is studied based on dark channel a priori.In order to solve the problem of inaccurate global atmospheric light estimation in sea fog image clearness application,a sea fog image clearness algorithm based on sky region segmentation is designed to make the atmospheric light be estimated more accurately by extracting the sky region,and the restored image quality get improved.In view of the problem of color distortion and halo in the sky,a new method for sea fog image clearness based on transmittance fusion is proposed.In the method,the atmospheric light is reestimated by mean of the quadtree segmentation,and the initial transmittance is estimated by constructing the cost function to avoid the color distortion and halo in the sky region,then in another way,the transmittance is estimated again to get more details by combining the Retinex theory,finally,the two transmittance images are fused and optimized to make the edges and details of the final transmittance get more accurate estimation.By this method,the halo effect in the salutatory areas of the scene depth is improved and the restoration effect of the sea fog image is better.Finally,the objective evaluation methods of image quality are introduced.In consideration of the special characteristics of sea fog images,some quality evaluation standards,including the image brightness,contrast,average gradient and halo intensity,are established to evaluate the restored images objectively.The experiment results demonstrate that the proposed algorithm is more effective in the sea fog image defogging,and can increase the contrast,sharpness and visibility of the image,enhance the details of the image and improve the quality of the sea fog image.