摘要
提出了一种基于低频系数局部区域梯度信息的多分辨率图像融合方法。根据局部梯度信息对源图像的小波低频系数进行选择,获取融合图像的对应低频系数。依照平均误差、峰值信噪比、均方根误差以及偏差度、熵等评价标准,将该方法的多聚焦图像融合效果与其他三种常用低频系数融合方法的效果进行了比较。实验结果表明,该方法获得的大部分评价指标都优于其他三种方法,且其最佳小波分解层数为2层,而其他三种方法的最佳小波分解层数为5层。最佳小波分解层数越少,图像融合的计算量越小。该方法在减少计算量的同时,提高了融合质量。
A multi-resolution image fusion method using local gradient of low frequency coefficients is proposed. According to local gradient information, wavelet low frequency coefficient obtained from source image is selected to compose respective low frequency coefficient of the fused image. The fused image by the proposed method is evaluated with some parameters such as average error, peak signal noise rate, root mean square error, difference index and entropy, in comparison with three other conventional methods. Experimental results show that most evaluation specifications acquired with this method are better than those with three other methods. The optimal wavelet decomposition level with this method consists of two levels, while that of the others consists of five levels. The less the optimal wavelet decomposition level is, the less the computational operation of image fusion will be. This method reduces the computational load and improves the quality of image fusion.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2005年第8期59-63,共5页
Opto-Electronic Engineering
关键词
图像融合
小波变换
分解层数
图像处理
Image fusion
Wavelet transform
Decomposition level
Image processing
