摘要
研究图像融合精度优化问题。针对传统的图像融合质量不高,由于图像中边缘信息丢失,使图像的清晰度和分辨率降低。为了提高图像的分辨率,给出了一种新的均值聚类多小波图像融合算法。针对多小波变换的优点,对融合图像采取分段式的多小波分解,并根据不同的融合规则对分解后的小波的不同高低频率重构,对融合后的边缘模糊图像区域采用均值聚类方法进行融合,可有效解决图像融合后边缘细节丢失等问题,对多组多聚焦图像进行验证实验,仿真结果表明,改进的方法能够更好的保留图像边缘信息,融合效果明显优于传统的图像融合方法。
The accuracy of image fusion optimization problem is studied.The traditional image fusion algorithm has the problems of low quality and low resolution.In order to improve the image resolution,the paper presented a new clustering algorithm for multi-wavelet image fusion.Making full use of multi-means clustering algorithm and the advantages of wavelet transform,the algorithm first carried out the wavelet decomposition of fusion images with different bands,and chose different fusion rules for the reconstitution of different frequencies.The mean clustering method was used to fuse the blurred image areas of the edges,which can effectively resolve the problem of edge detail losing after the,image fusion.The experiments of multiple sets and multi-focus images show that the improved method can keep a better edge information.The fusion effect is better than the traditional image fusion methods.
出处
《计算机仿真》
CSCD
北大核心
2011年第11期242-245,共4页
Computer Simulation
