Research On Polarization Image Dehazing Algorithm Based On Atmospheric Scattering Model

In foggy weather conditions,the quality of images usually drop dramatically due to the influence of atmospheric scattering particles,finding methods of improving image quality has always been an important research direction in the computer vision sector.The application of Polarized optical makes it possible for us to use the polarization information of the image.In this paper,based on the research of several algorithms of image enhancement and image restoration,an improved model for fog-degraded image is established in combination with the characteristics of polarized light and atmospheric scattering model.Under this model,the fog-degraded image can be recovered by solving the value of its transmittance and the value of light at infinity.The core of this algorithm is as follows: 1.Based on the dark grey transcendental theory,the value of infinite Atmospheric light intensity can be estimated under the constrain which is the multiplication result of the probability function of pixel points close to the sky and probability function of brightness limit is close to 1;2.The transmittance of fog-degraded image estimated by dark grey transcendental theory usually has the disadvantage of obvious regional effect,to optimize the transmittance of fog-degraded image,a bilateral filtering algorithm is improved by means of dimension reduction.In this paper,polarization images of three different polarization Angle(0 °,60 ° and 120 °)were gained by rotating polaroid under different fog conditions,In the algorithm validation phase,an objective evaluation which included image variance,average gradient,spatial frequency was established.By using this evaluation method,the fogdegraded images were compare with images processed by algorithm brought up in this paper,images processed by Retinex and images processed by dark grey transcendental theory.The experimental results show that the algorithm brought up in this paper can recover images.