无需重构的多分辨率图像融合算法(英文)  

A Multi-resolution Image Fusion Algorithm without Reconstruction

作　　者：王红梅[1] 韩治国[1] 陈励华[2] 

机构地区：[1]西北工业大学航天学院,西安710072 [2]第二炮兵工程学院301教研室,西安710025

出　　处：《光电工程》2011年第11期86-92,共7页Opto-Electronic Engineering

基　　金：西北工业大学基础研究基金(GBKY1008);西北工业大学"翱翔之星"人才计划项目;航天支撑基金项目;陕西省科技计划项目

摘　　要：提出了一种基于平稳小波变换的多聚焦图像融合算法。首先对待融合图像进行平稳小波分解,得到图像尺寸相同的低频分量和高频分量,然后对低频分量使用拉普拉斯能量进行清晰度判断,对于高频分量,则先计算其各个尺度,不同方向高频分量的绝对值和,进而通过能量特征判断其清晰度,最后通过比较低频分量和高频分量清晰度决策图的相同和相异性得到融合图像。计算机仿真实验表明,本文算法得到的融合图像清晰度较好,熵、平均梯度、空间频率和互信息等客观评价指标值高于平均法和传统基于小波变换的图像融合算法,互信息量比文献[3]中的方法提高了约2.4倍,是一种有效的多聚焦图像融合算法。A novel multi-focus image fusion algorithm is proposed on the basis of stationary wavelet transform without reconstruction. The original images to be fused are firstly decomposed by stationary wavelet transform. The low-frequency coefficients and high-frequency coefficients can be obtained. The clarity of low-frequency coefficients is estimated by the energy of Laplacian and the decision map can be acquired by comparing the energy of Laplacian. For the high-frequency coefficients at different resolution and different direction, the absolute value of these coefficients are summed firstly, and then the local region energy is used as the feature to get the high-frequency decision map. The fused image can be obtained by combining the low-frequency decision map and high-frequency decision map. The simulated results demonstrate that the proposed algorithm can get clearer image when compared with the average method, the traditional wavelet-based algorithm and some modified algorithm. The objective evaluation indices value, such as Shannon, average gradient, space frequency, standard deviation and mutual information, are improved compared with the average method and the traditional wavelet-based algorithm. Especially, the mutual information value computed by our algorithm is 2.4 times compared with the algorithm of Ref [3]. Subjective observation and objective evaluation suggest that our algorithm is effective for multi-focus image fusion.

关 键 词：图像融合 平稳小波变换 拉普拉斯能量 客观评价 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]