基于量子力学和拉普拉斯金字塔的图像融合方法  

Image Fusion Method Based on Quantum Mechanics and Laplacian Pyramid

作　　者：席亮[1,2] 

机构地区：[1]上海开放大学普陀分校信息工程系,上海200062 [2]上海市普陀区业余大学信息工程系,上海200062

出　　处：《计算机与现代化》2015年第4期37-41,共5页Computer and Modernization

基　　金：上海远程教育集团学科研究课题(JF1510)

摘　　要：针对数字图像的融合问题,研究一种基于量子力学理论的图像分解方法,并将其结合拉普拉斯金字塔变换,提出一种新的图像融合方法。先把2幅灰度图像表示成量子比特的形式,然后把每幅图像分解成4幅特征子图,根据分解后特征子图的不同含义,使用不同的融合策略分别对子图进行融合。其中一幅子图使用拉普拉斯金字塔变换进行融合,另外3幅子图使用区域梯度取大法进行融合,最后通过重构得到融合图像。仿真实验表明该方法融合效果较好,在主观视觉感受和客观评价指标上均优于传统的图像融合方法。A new image decompostion method based on quantum mechanics is proposed for the digital image fusion,combined with the Laplacian pyramid transform,a new image fusion method is proposed. Firstly two gray images are expressed as the form of quantum bit,and then each image is decomposed into four characteristic sub-images. According to the different meaning of the decomposed sub-images,different fusion rules are used. It uses Laplacian pyramid transform for one sub-image,and the method of selecting more local gradient is used for another sub-images. then the fusion image is obtained by reconstructed. The simulation results show that the method has better fusion effect,it outperforms the conventional image fusion methods in subjective vision perception and objective evaluation index.

关 键 词：量子力学 图像分解 拉普拉斯金字塔 区域梯度 图像融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]