暗原色雾天图像增强方法改进

An Enhancement Approach of Fog-Degraded Images Based on Dark Raw Color Algorithm

讨论图像增强技术中利用暗原色先验去雾算法的不足之处,并提出改进方法。提取一幅图像所有小区域中最暗点的坐标,得到其分布图;通过这些点得到较准确的透射率值,以此近似估计图中其他点的透射率,进而复原去雾后的图像。

The paper discusses the algorithm principle of utilizing dark channel prior to enhance the fog-degraded image and points out its limitations and methods for improvement.Through extracting the coordinates of the darkest points in all the micro-regions of the image,the distribution diagram of these points is formed.Based on these points,we can get the relatively accurate transmissivity value,through which we can estimate the transmissivity value of other points in the image,and then restore the image after defogging.