结合最小滤波和引导滤波的暗原色去雾

用基于暗原色信息的去雾方法处理雾化图像时,需要考虑消除景深边缘处的白色晕块并降低修复透射率过程所耗费的时间。为了既能够有效地去除白色晕块又能够降低修复透射率的处理时间,本文分析了最小滤波和引导滤波分别用于修复透射率的优缺点,提出了采用最小滤波和引导滤波相结合的方法来修复透射率。该方法综合了两者的优点,同时又弥补了各自的不足。较之于软抠图法,不仅有效消除了白色晕块,并且大大地降低了透射率处理时间,提高了处理效率。另外,针对传统暗原色去雾方法的缺点进行了改进。根据天空区域像素亮白的特性,将颜色和大气光相近的像素视为天空区域像素,然后对其进行颜色原值保留处理,从而有效地消除天空区域颜色的失真和跳变,从整体上改善了去雾复原图像的视觉效果。

Image dark channel prior haze removal based on minimum filtering and guided filtering

When a hazed image is processed by dark prior haze removal method, it should take the elimination of white halo on the depth-edge and the reduction of time consuming for modifying the transmittance into account. In order to eliminate the white halo on the depth-edge and to reduce the time consuming for modifying the transmittance, this article fully analyzes the advantages and disadvantages of minimum filtering and guide filtering in modifying transmittance, respectively. A new method in combination with the minimum filtering and the guide filtering is adopted to modify the transmittance, by which their advantages are taken and the inadequacies are made up. As compared with the traditional method, it not only eliminates the white halo but also reduces the processing time significantly. In addition, this paper improves the poor result of traditional dark prior method with a sky area in the hazy image. According to the bright-white features of the pixels in the sky area, the pixels with color values similar to that of the airlight are viewed as the sky area pixels. Then these pixels are processed to maintain their original color values and to eliminate the color distortion and jump existing in the sky area effectively. Thus, the visual effect of whole defogging restored image could be improved.