A Comparative Study-Dehazing Method Based on Homomorphic Filtering and Linear Transformation

Images captured via camera in the bad weather conditions like hazy or foggy ones would result in degradation of images. This is mainly due to the scattering of atmospheric light. This could lead to the deterioration of the image due to the scattering of atmospheric particles and due to the presence of haze. In this paper, a comparison between two algorithms is made, one is a fast algorithm for single image dehazing is proposed using the homomorphic filtering and the other is a single image dehazing using the linear transformation. In these both techniques, firstly, to accurately measure the atmospheric light an additional channel method is proposed based on quad-tree sub-division. After finding the atmospheric light portion in the input image filtering is being applied. For both the techniques, filtering is being applied, in which homomorphic filter is used in one and in the other minimum filtering is being used and then gaussian blurring. At last the haze-free image is obtained as the result of atmospheric scattering model. Number of both subjective and objective evaluations are being done with the aid of experiments which shows that this algorithm can clearly and naturally recover the image, with at most clarity at the edges where there is sudden changes in the depth of field. Here the algorithmic time complexity is the linear function of the image size. This has advantages in the running time my maintaining balance between the running time and processing.