一种改进的多聚焦图像融合算法  

An Improved Multi-Focus Image Fusion Algorithm

作　　者：郭德全[1] 巨辉[1] 

机构地区：[1]成都信息工程学院,四川成都610225

出　　处：《计算机仿真》2009年第11期251-254,共4页Computer Simulation

摘　　要：针对处理多传感器采样到的聚焦点不同的图像,高、低频融合规则的选择问题,提出一种改进的基于小波变换多聚焦融合算法。根据小波变换后的低频分量和高频分量的不同特性,算法分别选用改进能量算法和均匀测试度的方法进行融合,再通过小波逆变换得到融合后的图像。实验结果表明,由于充分考虑了小波变换的特点与人眼系统的特性,与其它融合方法得到的图像相比,提高了融合图像包含的信息量,较好地消除了局部对比度极性反转的情况,明显地增强了融合图像的清晰度,而且保留了源图像中的边缘细节,得到了较好的融合效果。The article introduces an improved multi -focus image algorithm based on wavelet transform, which is used to fuse multi - focus image from different sensors, with an aim at the selection of high and low frequency fusion rules. According to the different trait of coefficients of low frequency and high frequency, this algorithm chooses an improved energy algorithm and average test degree rule to fuse low and high frequency of the image , then obtains the fused image by wavelet inverse transform. Compared with other fused images obtained by other algorithms, because the algorithm considers fully the characteristics of wavelet transform and features of human vision systems, it evidently improves the inclusive information of image fusion, eliminates the circum stances of local contrast polarity reversal, tones the definition of image fusion and commendably maintains the edge detail of input images, and gains a better fusion effect.

关 键 词：多聚焦图像 小波变换 局部区域能量 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]