暗原色雾天图像增强方法改进

摘要

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 val-ue of other points in the image,and then restore the image after defogging.%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.