摘要
以SAR与可见光图像、多光谱与全色图像为研究对象,提出了一种基于曲波变换的遥感图像融合方法。首先将图像进行曲波变换,然后在不同的频率域利用融合规则融合曲波系数,最后通过重构得到融合图像。采用均方误差、偏差指数等指标对融合效果进行了客观评价,并与基于小波变换的融合进行了比较。实验结果表明,该方法在保留原始图像重要信息、抑制噪声能力方面均优于小波变换方法。
Compared with wavelet,Curvelet has much better identification ability of directions, so it is more appropriate for the analysis of the image edges such as curve or line characters than wavelet. When introducing curvelet transform to remote sensing image fusion,we can take the characteristics of original images well ,The ability of noise suppression is also better than wavelet transform. So the method of curvelet transform based image fusion is proposed. Firstly,the original images are decomposed using curvelet transform,then the coefficients are fused with the different fusion rules in the different frequency bands, finally the fused coefficients are reconstructed to obtain fusion results. Mean square error, difference coefficient etc. are used to evaluate the results. The fusion images are compared with that based wavelet transform The results show that this method can get much better fusion results than wavelet in two respects：preserving original images＇ important information and noise suppression.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2008年第3期400-403,411,共5页
Journal of Optoelectronics·Laser
基金
国家自然科学基金资助项目(60675015)
关键词
图像融合
曲波变换
小波变换
遥感图像
image fusion
curvelet transform
wavelet transform
remote sensing image
