基于形态学4子带分解金字塔的图像融合

Image Fusion Based on Morphological 4-Subband Decomposition Pyramid

提出了一种基于数学形态学滤波的多分辨力图像融合。这种融合方法使用了形态学开闭运算构造了低通与高通滤波器,将原始图像分解为4子带图像金字塔和4子带方向衬比度图像金字塔。然后利用方向衬比度和区域标准差进行图像融合得到融合的4子带图像金字塔,最后应用子带图像重构得到融合图像。融合实验表明,该方法优于传统的形态学金字塔图像融合,衬比度金字塔图像融合和小波分解图像融合。

A novel multi-resolution image fusion based on mathematical morphological filtering is proposed. This fusion scheme constructs the low-pass and high-pass filters with morphological opening and closing operators. The original image is then decomposed into 4-subband image pyramid and 4-subband direction contrast image pyramid. The image fusion is performed by the direction contrast and region standard deviation to get the fused 4-subband image pyramid. The final fused image is formed with the subband image reconstruction. The fusion experiments indicate this image fusion is superior to the traditional fusion based on the morphological pyramid, contrast pyramid and wavelet decomposition.