一种基于区域生长的红外与可见光的图像融合方法

针对目前红外与可见光目标跟踪的视频序列融合难以满足实时性问题，利用红外图像目标与背景显著的灰度差异特征，结合目标跟踪中目标分割时常用到的区域生长法，通过区域生长方法从红外图像中提取目标区域，再将得到的红外目标区域与已经过图像配准的可见光图像的背景区域进行融合处理，最终得到既具有红外图像较好的目标指示特性又具有可见光图像清晰场景信息的融合图像。实验表明：该算法不仅简单易行，而且所得到的融合图像视觉效果优于其他融合算法得到的图像。

Fusion of IR and visible images using region growing

In order to meet the real-time requirement of the video sequence fusion in visible and infrared target tracking,region growing method,a frequently used method in target segmentation during target tracking;was used to extract the target regions from IR image through the gray difference between IR targets and its backgrounds.The extracted target regions of IR image are fused with background regions of visible image.The fused mosaic image has good IR target features as well as clear visible background.Experiment shows that the method is simple and effective.