Research On Sea-sky Video Image Defogging Based On Dark Primary Colors

In the foggy environment under the sea and sky background,the video image quality obtained by the computer vision system has different degrees of attenuation,resulting in the loss of video image information,which will cause great interference in target recognition,target tracking,target feature extraction and other related fields.Therefore,it has important practical significance and application background to remove fog from sea-sky video image and restore clear sea-sky video image.The purpose of this paper is to analyze and improve the problem of dark primary prior defogging algorithm in sea-sky video image defogging,and to verify the effectiveness of the improved algorithm through related experiments.The specific work is as follows:(1)The dark priori defogging algorithm is used to defog images in different sea-sky scenarios.The related experiments are carried out to analyze the defogging images,and the problems of the dark priori defogging algorithm in the sea-sky background are summarized.(2)In this paper,the problem of false selection of atmospheric light value in the sea and sky background is analyzed,and the problem of false selection of atmospheric light value is solved by using image quartering method,and the effectiveness of this method is verified by relevant experiments.This paper analyses the causes of color bias in the sky area of sea-sky defogging image,and improves the tolerance method by fitting the correlation function through a large number of experiments.The time-consuming problem of dark primary color prior defogging algorithm is analyzed,and the efficiency of dark primary color prior defogging algorithm is improved by using guide filter algorithm instead of matting algorithm and up-down sampling method.The effectiveness of this method is verified by relevant experiments.By synthesizing the above improved methods,the fog removal of fog simulation images in different sea-sky scenarios is compared with other fog removal algorithms.The experimental results verify the effectiveness of the improved dark priori fog removal algorithm.(3)Based on the sea-sky image defogging algorithm,this paper mainly studies the sea-sky video image defogging.The atmospheric light values of video images in different sea-sky scenes are counted under different color channels,and the method of selecting atmospheric light values of sea-sky video images is determined in this paper.That is to say,the atmospheric light values are calculated repeatedly every ten seconds.Considering the redundancy between frames,the method of selecting the transmittance map of sea-sky video image in this paper is determined by relevant experiments,that is,comparing the similarity between adjacent frames,deciding whether to transfer the rough projection map of the previous frame to the next frame according to the similarity size,and then using the gray-scale map of each frame to obtain the fine transmittance map.Finally,combined with the above selection method of atmospheric light value and transmittance map of sea-sky video image,the sea-sky video image and fog simulation video image in different scenes are defogged,and compared with the method of frame-by-frame defogging of video.Through relevant experiments,the effectiveness of the defogging algorithm of sea-sky video image in this paper is verified,that is,the quality of defogging of sea-sky video image is guaranteed.It also improves the efficiency of fog removal.