基于局部色调映射的异源图像匹配方法

色调映射是一种适用于亮度和对比度变化的快速图像匹配方法。由于异源图像间存在复杂的灰度变换关系,直接采用色调映射方法进行匹配的成功率通常难以满足应用需求。为了提高匹配成功率,本文提出基于局部色调映射的异源图像匹配方法。首先将实时图划分为不重叠的子区,对每个子区进行直方图均衡化和弱切片变换,通过局部色调映射计算子区与基准图的距离系数图。融合全部距离系数图得到图像匹配结果。实验结果表明,该方法匹配成功率高于现有的色调映射方法,且计算时间仅略高于现有方法,优于异源图像匹配中常用的互信息方法。

Matching Multi-sensor Images Based on Local Tone Mapping

Tone mapping is an efficient image matching algorithm adaptive to brightness and contrast variation. Due to the complex and nonconforming relation between multi-sensor images, tone mapping can hardly be directly used on multi-sensor images with satisfactory result. To improve the matching rate, the local tone mapping algorithm is proposed which includes the following steps. The template map is divided into nonoverlapping blocks. Each block is processed through histogram equalization and weak slice transform. Distance maps are calculated for every block by local tone mapping. The matching result is obtained by fusion all these distance maps. Experimental results show that, the matching rate of local tone mapping significantly outperforms tone mapping and mutual information, while its computation time is slightly longer than tone mapping and less than mutual information.