摘要
提出了一种新的基于小波变换多尺度积局部区域统计量的图像融合算法,简称MPLVDDWT(multiscale product local variance of dyadic discrete wavelet transform)算法.在图像融合过程通过利用多尺度积从而隐含了一个去噪的过程,这有利于在融合图像中突出图像的细节特征.利用统计分析的评判准则,如熵、标准偏差评价图像的融合效果.实验结果表明,该方法提高了图像的熵和标准偏差.在保留原图像信息的情况下增强融合图像的细节信息.
A novel image fusion algorithm named MPLVDDWT （multiscale product local variance of dyadic discrete wavelet transform） is proposed which is based products of wavelet transform. A denoising process is hided benefits to enhance detail features in image fusion. Entropy and on the local statistic of the multiscale by using multiscale products, which standard deviation are used to evaluate image fusion performance. The experimental results show that the image fusion effect is improved by the method.
出处
《测试技术学报》
2007年第5期428-433,共6页
Journal of Test and Measurement Technology
基金
国家自然科学基金重点资助项目(10234070)
关键词
小波变换
多尺度积
图像融合
局部统计量
熵
标准偏并
wavelet transform
multiscale product
image fusion
local statistic
entropy
standard deviation
