一种改进的可见光与红外图像融合算法

Improved Fusion Algorithm for Infrared and Visual Images

针对红外和可见光图像的特点,提出一种基于加权对比度和局部熵相结合的多分辨图像融合算法。利用小波变换得到待融合图像的多分辨结构,在图像多分辨结构相应各级上采取不同的融合规则,得到融合图像的多分辨结构,利用小波逆变换重构融合图像。实验结果表明,该方法在保留可见光图像高的空间分辨率和丰富的纹理细节信息的同时,融合了红外图像对热目标比较敏感的特点,提高了图像的可判读性。

Aim to the features of infrared and visual images, a fusion algorithm based on contrast and local entropy is proposed. For the multi-resolution structure of wavelet transform, different decomposing levels use different fusion rules. The fusion image is reset through wavelet transform. The results show that the proposed method can keep the spatial resolution of visual images and texture detail information of infrared images, and enhances the readability of source images.