Research On Single Image Dehazing Algorithm Based On Four Fork Tree And Block Region

Picture is an important way for people to get information.Heavy air pollution is caused by industrial emissions,burning of straw and untreated coal heating.The number of haze weather days is increasing.The fog in the photos taken in the outdoor environment affects the observation of the focus of the target,which greatly reduces the visual effect.Therefore,it is a very urgent need to transform the fog map into high definition image.After a lot of researchers have done a lot of research in the domain of image sharpening algorithm,the effect of the image demogging algorithm has reached a good level.At present,in all the fogging algorithms,the He DCP algorithm is the best for fog removal,but it is found that the algorithm appears discontinuous regions and colors when there are large areas of fog like area in the image processing,and the halo effect will appear in the area where the depth of the scene is abrupt in the edge of the sky area.Because human eyes are more sensitive to the distortion of image edges and contours,the visual effect of halo effect will be very bad.At the same time,the processing speed is difficult to achieve real-time processing requirements,which seriously limits the practical application of the algorithm.In view of these problems,this paper first studies the degradation mechanism of fog images,analyzes the advantages and disadvantages of the existing algorithms,and makes the following major improvements:(1)First of all,the foggy image obtained from the image acquisition device is sampled in Pyramid,which is reduced to 1/4 of the fog graph,and the reduced fog image is fogged.Finally,the linear interpolation algorithm is used to restore the size of the original image.In the DCP algorithm,the A(A is the estimated atmospheric light value)is the pixel average of the first 1% points in the original pixel,thus the A value of each color channel is likely to approximate the maximum of the gray level,which will cause the color deviation of the processed image too much and there are many block areas and unconnected images in the sky surface.In this case,the Canny algorithm is used to segment the image,to obtain the atmospheric light and transmittance for different block regions,to refine the transmittance by the guide filter,and to optimize the transmittance with the adjustment factor,and to compare the gray mean values of different block regions recursively by four fork tree,which can significantly reduce the execution of the algorithm.(2)In view of the effect of atmospheric light A on fogging effect,an improved scheme is proposed in this chapter.According to the ratio of One-Dimension Entropy and the size of the gray mean value of the original image,the original image is divided into several different classes,and then the appropriate proportion is selected to set the A values of each classification.The effect that appears after the dehazing becomes clearer.