基于拉普拉斯金字塔改进的图像融合方法

Improved Image Fusion Method Based on Laplacian Pyramid Transform

改进传统的拉普拉斯金字塔融合图像方法,对源图像进行一系列层层递进的处理,而非传统算法直接对源图像进行拉普拉斯变换,与经过处理后的特征更明显的图像作变换。先提取图像的轮廓特征,然后对比得到的两幅梯度图像间对应的像素值,找出每幅图像比另一幅图像清楚的区域,保留各自的优势部分,据此建立两幅梯度图像的二值图像,进行拉普拉斯金字塔变换的源图像就是上述处理获得的二值图像。接着融合变换后的两个图像金字塔,最后,需要与最初源图像的高斯金字塔重新构建图像,将前一步获得的拥有更多高频细节的拉普拉斯金字塔从塔顶开始逆变换,直至变换出塔底图像为止,即可得到最终结果。实验数据显示,改进的融合方法起到了持有更多细节的作用。

An improved image fusion algorithm based on Laplacian pyramid transform is presented in this paper. First, use gradient to get the feature extraction of the source image, which is called gradient image, then compared each pixel value of the two gradient image to confirm the clear part of each image, create the binary image of the gradient image to keep the clear part of each image, this binary image is the source image which use Laplacian pyramid transform to decompose. Decompose the binary image to obtain layers of Laplacian pyramid images. For different layers of decomposition images, different fusion method should be used. Finaly, use the fused Laplacian pyramid and the Gaussian pyramid of the source image to reconstruct from the top layer by layer and obtain the fused image.