基于NSCT和边缘检测的图像融合算法研究

近年来,图像融合已经成为计算机视觉领域一向有价值的新技术,论述了NSCT变换理论,先将图像作非下采样拉普拉斯金字塔尺度分解,在各个尺度层对高频子带作非下采样方向分解,然后分别采用基于区域能量和边缘检测的融合规则得到融合图像的非下采样Contourlet低频和高频系数;最后再进行非下采样Contourlet逆变换得到融合图像。实验结果表明,该方法的融合图像多项指标都优于Contourlet变换、NSCT变换。

Image Fusion Based on Nonsubsampled Contourlet Transform and Edge Detection

In recent years, image fusion has become a valuable and new technology in computer vision field. This paper discusses the theory of Nonsubsampled Contourlet transform. First, multiscale decomposition of the source images is performed with nonsubsampled Laplacian pyramid, and directional decomposition of the high frequency subbands in every scale is also done with nonsubsampled directional decomposition. Then, the low frequency coefficients and high frequency coefficients for the fused image could be obtained by means of the fusion rule respectively based on local energy and edge detection. Finally, the fused image is reconstructed by the inverse NSCT. Experimental results show that the proposed method could work better than Contourlet transform and NSCT.