Image Haze Removal Method Based On Dark Channel Prior And Sky Region Detection

Image defogging is a technology to enhance fogged images or to restore fog-free scenes,to show the information of the image in a better way,so as to get better visual qualities of an image or better subsequent analysis and treatment in computer system.In real life,defogging system can be widely used in many fields such as military,video monitoring,satellite imaging,and is of great value and significance.Due to the complexity of the defogging,currently few algorithm can solve the problem perfectly.The pursuit of faster and better defogging effect has always been the goal of scientific research workers.This paper first introduces the atmospheric scattering model,which can be expressed as a physical model and is widely used to explain the process of image degradation and formation of a fogged image.Then is detailed analysis of defogging application of dark channel prior principle,analysis of the deficiencies in the processing speed and processing effect in the sky area.Dark channel prior method often gives a smaller value in transmission estimation of sky region,which leads to serious color distortion in the image with uneven color block after haze removal.To solve these problems,this paper proposes an improved single image haze removal method based on sky region detection,which uses quadtrees model in the sky area to get accurate atmospheric lighting value and total variation regularization to optimized transmission diagram.This method can not only retain the good results of dark channel prior method but also solve the problem of dark channel prior method which can't apply in sky region.Besides,this paper also proposes a new and simple method of image fusion,which eliminate the obviously white border of the fused image.In view of the fact that the fog weather image segmentation is not significant,the Laplacian variance operator is used to determine the concentration of fog,and the image is processed by the improved tolerance mechanism under the fog condition.The results show that this method can deal with the color distortion and the color block of sky region which achieves good results in final restored image.