激光与光电子学进展, 2022, 59 (4): 0411001, 网络出版: 2022-01-25
四分暗通道均值比较法的双角度偏振图像去雾 下载: 782次
Dehazing of Dual Angle Polarization Image Based on Mean Comparison of Quartering Dark Channels
摘要
雾天环境下拍摄图像时,雾、霾等介质导致图像模糊、对比度低、色彩暗淡。结合大气散射模型与偏振光原理,实现了利用斯托克斯参数获取亮度最大和亮度最小的双角度偏振图像,并通过新提出的四分图像暗通道均值比较法,准确估计出场景中无穷远处大气光强值,最终将有雾图像恢复到无雾图像。所提方法在薄雾和浓雾环境下均取得了较好的实验结果,在两种环境下图像NRSS(no reference structural similarity)和平均梯度均有较大提升。
Abstract
When taking images in foggy environment, fog, haze, and other media lead to image blur, low contrast, and dim color. In this paper, the atmospheric scattering model and the principle of polarized light are combined to obtain the dual angle polarized image with the maximum brightness and the minimum brightness by using the Stokes parameters. The atmospheric light intensity at infinity in the scene is accurately estimated by the new method of comparing the mean value of the dark channel of the quartered image, and finally the foggy image is restored to the image without fog. The experimental results show that the image NRSS (no reference structural similarity) and mean gradient are greatly improved in both mist and dense haze environments.
孟宇飞, 王晓玲, 刘畅, 周哲海. 四分暗通道均值比较法的双角度偏振图像去雾[J]. 激光与光电子学进展, 2022, 59(4): 0411001. Yufei Meng, Xiaoling Wang, Chang Liu, Zhehai Zhou. Dehazing of Dual Angle Polarization Image Based on Mean Comparison of Quartering Dark Channels[J]. Laser & Optoelectronics Progress, 2022, 59(4): 0411001.