The Research Of Image Defogging Algorithm Based On Dark Channel Prior

With the rapid development of computer science and digital image processing technology,digital image defog technology has played a very important role in the field of computer vision.In recent years,large-scale fog and haze weather appeared frequently,the images collected outdoors will be severely degraded by the scattering of suspended particles in the atmosphere,in this case,only rely on imaging device upgrade cannot obtain an outdoor image effectively,so the vital problem is how to use the information of the single fog image to restore the better visual image.Based on the dark channel prior algorithm the specific research content of this paper is as follows:Firstly,according to the problem that the dark channel prior algorithm has the color distortion after processing the high brightness image regions such as the sky,this paper propose a new variational energy model including a term of dark channel prior algorithm,a data fidelity term and a gradient extraction term.According to the transmittance,the energy model assigns the weight of the first two items.It makes the color of the scene image closer to the original image in the field with larger depth,and it keeps a better fog removal effect in the field with smaller depth.Then,the energy functional problem can be solved by the gradient descending flow method,and use the finite iteration of the iterative equation to obtain the image without fog.The experimental results show that the algorithm works well for solve the problem that the dark channel prior algorithm has a bad effect on the image regions such as the sky,and finally achieves better fog removal effect.Then,according to the block effect of the minimum value filtered transmittance image in the prior algorithm of dark channel,and the effect of single scale guided filtering fog removal is not ideal,this paper propose an optimization algorithm,which combining multiscale minimum value filtering and multi-scale guided filtering.In this scheme,high quality transmittance images are obtained by combining the weighted images of different scales after filtering.The experimental results showed that the optimization algorithm has better fog removal effect,while a more natural and real color defog image is obtained.