基于非子采样Contourlet变换的图像融合算法

Image Fusion Algorithm Based on the Nonsubsampled

提出了一种基于非子采样Contourlet变换的图像融合算法.首先,对图像作Contourlet变换,得到一个低频子带和一系列高频方向子带.其次,对低频子带系数采用基于边缘的规则来融合,而对于高频方向子带系数,采用基于一种基于Contourlet局部系数相关性的规则来进行融合.最后,对融合系数进行非子采样Contourlet逆变换,得到最终的融合图像.实验结果表明,该方法对可见光、红外图像融合十分有效,能较好地提取原始图像中的边缘方向信息和几何特征.

An image fusion algorithm based on Nonsubsampled Contourlet Transform （NSCT） is proposed. First, the Contourlet transform of the source images is performed to get a low-frequency sub-band and a series of high-frequency subbands. Then, the low-frequency sub-band coefficients of the fused image can be obtained by means of the fusion hale based on the edge, while the high-frequency sub-bands coefficients can be combined by using the fusion rule based on the relevance of the Contourlet coefficients. Finally, the final fusion image has been achieved by implementing the inverse NSCT of the fused coefficients. The results indicate that the algorithm is effective for the fusion of the visible and infrared images, which can better extract the edge directional information and the geometric characteristics.