资源描述：
《多维小波变换工具箱》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、多维小波变换工具箱Multi-dimensionWaveletToolboxVersion1.0作者：陈甫一、函数的简要说明convnd对多维数据在某一维向上做线性卷积convndcut和convnd差不多，但加上了可以截取卷积结果的中间一部分而把两边删去的功能dyaddownnd隔点采样dyadupnd在两点间添0dwtnd1多维空间中一个维向上的一次小波变换dwtndn多维空间一次小波变换(各维向都做一次)wavedecnd多维空间多次小波变换分解idwtnd1多维空间中一个维向上的一次小波
2、逆变换idwtndn多维空间一次小波逆变换waverecnd多维空间多次小波逆变换重构packup利用dwtndn和wavedecnd的结果拼成多维系数空间的表示形式fitscale改变数据值的分布以利于图像显示show3d用图像显示三维空间的小波变换系数二、实例实例一dwtnd1和idwtnd1r=rand(4,4,4)%随机生成一个多维空间的数据，这里是个4*4*4的三维空间r(:,:,1)=0.20280.19880.93180.52520.19870.01530.46600.20260.
3、60380.74680.41860.67210.27220.44510.84620.8381r(:,:,2)=0.01960.50280.18970.54170.68130.70950.19340.15090.37950.42890.68220.69790.83180.30460.30280.3784r(:,:,3)=0.86000.89980.66020.53410.85370.82160.34200.72710.59360.64490.28970.30930.49660.81800.3412
4、0.8385r(:,:,4)=0.56810.44490.95680.97970.37040.69460.52260.27140.70270.62130.88010.25230.54660.79480.17300.8757[ca,cd]=dwtnd1(r,1,'haar')%对数据r在第一维向即row方向做一次haar小波分解ca(:,:,1)=0.28390.15140.98840.51460.61940.84280.89441.0679ca(:,:,2)=0.49560.85720.2709
5、0.48970.85650.51870.69650.7610ca(:,:,3)=1.21171.21720.70870.89180.77081.03440.44610.8116ca(:,:,4)=0.66360.80571.04610.88470.88341.00140.74470.7977cd(:,:,1)=0.00290.12980.32940.22800.23450.2133-0.3023-0.1174cd(:,:,2)=-0.4678-0.1461-0.00270.2763-0.3198
6、0.08790.26830.2259cd(:,:,3)=0.00450.05530.2250-0.13650.0686-0.1224-0.0364-0.3742cd(:,:,4)=0.1398-0.17660.30710.50080.1104-0.12270.5001-0.4408这里得到ca，cd分别是小波分解后的近似分量和细节分量，都是2*4*4的三维空间。q=idwtnd1(ca,cd,1,'haar')%用分解的结果做一次逆变换来得到原来的数据q(:,:,1)=0.20280.19880
7、.93180.52520.19870.01530.46600.20260.60380.74680.41860.67210.27220.44510.84620.8381q(:,:,2)=0.01960.50280.18970.54170.68130.70950.19340.15090.37950.42890.68220.69790.83180.30460.30280.3784q(:,:,3)=0.86000.89980.66020.53410.85370.82160.34200.72710.593
8、60.64490.28970.30930.49660.81800.34120.8385q(:,:,4)=0.56810.44490.95680.97970.37040.69460.52260.27140.70270.62130.88010.25230.54660.79480.17300.8757可以看到和原来的数据是一致的。实例二dwtndn和idwtndn还用刚才生成的数据ry=dwtndn(r,'haar')%对三个方向都做小波变换y=Columns1through4[2x2x2double
