Color Image Defogging Method Based On HSI Space And Wavelet Transform

Fog is a common weather phenomenon in nature,The images taken in in foggy weather will have degradation phenomenon and the image definition will be relatively low,These degraded images will have a great impact on traffic,space,military and other fields.Therefore,it is important for the study of image defogging technology.Recently defogging methods are of two main types: image restoration and image enhancement,and the image enhancement method is more widely used.This paper is through used image enhancement technology to defog the foggy image.Firstly,This paper introduces fog formation and foggy image characteristics;Then the HSI color space,wavelet transform and Retinex algorithm are analyzed;Finally,using HSI intensity model,wavelet transform and Retinex algorithm to defog the image.This paper also do contrast tests,and analyze the advantages and disadvantages of these algorithms based on the experimental results.This paper improve the intensity model of HSI color space and the traditional Retinex algorithm on these analysis:(1)Through the contrast tests of HSI space light intensity model,it is found that the fogging effect is unsatisfactory,and the image image is dark and not clear.Therefore,this paper use the improve intensity model of HSI color space to change the intensity of the image,Objective to improve the intensity of the image.Then deal the intensity image by wavelet transform.Not only the image intensity is improved,but also after wavelet transform,the image is divided into high frequency component and low frequency component,we can be treated separately,it can maintain image detail while enhancing image contrast.(2)The traditional Retinex algorithm to get the image with clear fog halo artifacts phenomenon,so this paper improved the traditional Retinex algorithm,we lead into discrete cosine transform based on traditional Retinex algorithm,it makes the defogging image to avoid halo artifacts.Through the contrast tests,our algorithm have obvious effect in defogging image.