针对雾霾天气下图像信息模糊难辨的问题,以图像的颜色特征传递为基础,提出一种基于自适应分类颜色传递的两次图像清晰化校正算法.首先采用基于统计特性的颜色传递方法对源图像进行一次整体校正;其次使用Kohonen神经网络对一次校正后的源图像和目标图像进行颜色信息的归类,使用归类后最相近的图像块进行二次颜色信息校正.通过两次颜色信息校正,图像整体信息和细节信息都得到了较完整和自然的恢复,效果良好.%Based on the color information transferring,it proposes a new algorithm to resolve the clearing problem for the pictures in the haze days'condition.This algorithm consists of two times clearing steps which use the selfadapt color categorize algorithm.The first step is for correcting the picture whole once,the second step is for categorizing the color and transferring the color information between the categorized color blocks.It uses Kohonen neural network to categorize color in the source & target pictures.This algorithm can correct ideally the pictures in haze days condition.
展开▼
