基于半反图像的导向滤波光补偿快速去雾算法

针对传统的基于暗原色先验的单幅图像去雾算法速度慢、效率较低,去雾不彻底以及在明亮区域颜色失真等问题,提出了一种基于半反图像的导向滤波光补偿快速去雾算法,将经过阈值调整的半反图像作为导向图,通过引导滤波得到大气散耗图,根据大气散射物理模型恢复清晰图像,并将恢复的图像进行参考白光补偿得到最终的去雾图像.实验结果表明:算法不仅降低了传统算法的时间复杂度,而且有效恢复了场景的对比度和饱和度,近景去雾彻底,远景去雾效果提升明显.

Fast light compensation defogging algorithm based on semi-inverse image guided filtering

Fast light compensation defogging algorithm based on semi-inverse image guided filtering is proposed to solve problems such as slow speed,low efficiency,inexhaustive defogging,and color distortion in bright areas of the traditional single image defogging algorithm which is based on dark channel prior.This algorithm will take the threshold value adjusted semi-inverse image as the guide image.By guiding the filter wave,a atmospheric veil is acquired,based on the physical model of which to restore the clear image.The restored image will be compensated with reference white light to get the final defogging image.The experimental result shows that,this algorithm not only has lowered the time complexity of the traditional algorithm,but has regained the contrast and saturation level of the scene effectively,achieving thorough defogging in the close view,and significantly improve defogging in the distant view.