The image fusion rules are studied according to the characters of wavelet transform. The wavelet di-rectional derivative(DD) and direction engineering(DE) for source images as well as wavelet feature of image texture and contour are defined, and then a new image fusion algorithm is proposed. The fusion criterion of al-gorithm is effective and the structure is simple. The fusion experiment of remote sensing and multi-focus images show that the algorithm can integrate the information and retain the features of source images more effectively and faster, with higher objective evaluation compared with several other excellent algorithms in the field of multi-spectral image fusion.%针对小波变换的特点,研究了图像融合规则,定义了图像小波方向导数和方向能量、图像纹理和轮廓的小波特征度,并提出了一种全新的图像快速融合算法.算法融合判据有效,结构设计简单.通过对多聚焦和遥感图像融合实验表明:该算法能有效地融合源图像的信息,保持源图像的纹理和轮廓特征.与同类优秀算法相比,主观评价效果相近,客观评价效果更高,运算时间更短.
展开▼