基于大气面纱优化和透射率修正的图像去雾

摘要

The image dehazing method based on median filtering can't effectively remove the haze of tiny edges and sky areas are prone to distortion. So, a single image dehazing based on the atmospheric veil optimization and transmission cor-rection is proposed. First it extracts sky area by using threshold segmentation and reduces gray value of corresponding region of the minimum image of each color channel, and then gets more accurate atmospheric veil by using weighted guided filtering, so that gets more accurate transmission, finally restored image is obtained through the atmospheric scattering model. Compared with other algorithms, this algorithm optimizes the atmospheric veil and correct transmission at the same time, the dehazing effect is more obvious in the tiny edge, and can solve the problem of distortion of the color of the sky area.%基于中值滤波的图像去雾算法不能有效去除细小边缘处的雾,且天空区域容易出现颜色失真问题.因此,提出了一种基于大气面纱优化和透射率修正的单幅图像去雾算法.首先基于阈值分割提取天空区域,并降低各颜色通道最小值图像中对应区域的灰度值,然后使用加权引导滤波得到优化后的大气面纱,从而得到较为准确的透射率,最后由大气散射模型得到复原图像.相比于其他算法,该算法在优化大气面纱的同时修正了透射率,在边缘处的去雾效果更加明显,同时解决了天空区域的色彩失真问题.