动态大气散射系数的颜色衰减先验图像去雾

摘要

针对颜色衰减先验图像去雾算法存在对较浓的有雾图像去雾效果不佳的问题,提出基于动态大气散射系数的颜色衰减先验图像去雾算法.用动态大气散射系数取代颜色衰减先验去雾算法中恒定大气散射系数的假设,定义大气散射系数为关于图像景深的指数函数.利用Middlebury stereo datasets中无雾图像和相应的景深图像得到合成有雾图像.采用均方误差(MSE)和结构相似度(SSIM)的综合评价参数MSE-SSIM确定上述指数函数的两个参数的最佳取值.实验结果表明与颜色衰减先验算法、He、Meng算法相比,该算法的去雾图像清晰颜色自然,有效地提高了去雾效果.%Aiming at the poor effect of the image haze removal algorithm using color attenuation prior to processing the medium haze image,a novel method is proposed that a image haze removal algorithm using color attenuation prior based on dynamic scattering coefficient of the atmosphere in the paper.In this algorithm,using dynanmic scattering coefficient of the atmosphere replaces the constant scattering coefficient of the atmosphere in the image haze removal algorithm using color attenuation prior,by defining the exponential function of the scattering coefficient of the atmosphere on the depth of image.Using ground truth images and corresponding depth images in Middlebury stereo datasets obtains synthetic hazed images.Adopting the MSE-SSIM comprehensive evaluation parameter of mean square error (MSE) and structural similarity index measurement (SSIM) determines the optimal value of the two parameters of the exponential function.Experimental results show that the dehazed image of proposed algorithm is clear and natural than the image haze removal algorithm using color attenuation prior,He and Meng.The proposed algorithm can effectively improve the effect of dehazing.