【摘要】 雾气环境下获取的图像往往对比度低、色彩弱化,严重影响了其后续应用。已有的去雾算法存在运算速度慢、色彩失真等问题。该文对基于暗原色先验的图像去雾算法进行了改进。(1)通过边缘检测得到图像边缘和非边缘区域,并分别采用不同大小的窗口计算透射率,提高透射率计算速度;(2)通过采用自适应调节的透射率下限值解决不满足暗原色条件的区域的色彩失真问题。实验结果表明,该文方法可以有效的解决图像局部色彩失真问题,大大减少运算时间,使复原图像整体视觉效果增强。更多还原

【Abstract】 The images acquired under mist environments are often low-contrast and weak colored,which seriously affects the subsequent applications.The existing dehazing algorithms have some issues,such as slow computing speed and color distortion.In this paper,an improved dehazing algorithm based on the theory of dark channel prior is proposed.Firstly,the image edge and non-edge regions are obtained through edge detection.Furthermore,the transmittance map is calculated across windows with different sizes on each region,which leads to the improvement of the computing speed.Secondly,by using adaptive lower limit of transmittance,the color distortion issue occurring in regions where the dark channel prior is not well satisfied(such as sky,water and other white and bright objects areas),is improved.Experiment results demonstrate that the proposed method can effectively solve the issue of local image color distortion,greatly reduce the computation time and enhance the overall visual effects of recovered images.更多还原