Matlab-EMD工具箱

EMD工具箱,在MATLAB中實(shí)現(xiàn)EMD的各種功能所必須的工具箱。

工具箱的安裝

運(yùn)行install_emd.m文件可以實(shí)現(xiàn)此工具箱的安裝，uninstall_emd.m實(shí)現(xiàn)卸載。

1、首先下載emd工具箱，50樓網(wǎng)址里面可以下。下載后解壓放在matlab的work工作路徑下package_emd文件夾。

2、打開(kāi)matlab，選擇File- Set Path- Add with Subfolders-你剛才下載的工具箱（package_emd）點(diǎn)進(jìn)去- Save- Close。

3、此時(shí)選擇work下package_emd文件夾作為工作路徑，即是C:\Program Files\MATLAB\R2010a\toolbox\package_emd。

4、在Command Window里面輸入mex -setup回車(chē)，問(wèn)是否選擇已有的編譯器你選y回車(chē)，再問(wèn)選擇哪個(gè)編譯器，你可以選擇C++的那個(gè)選擇相應(yīng)的編號(hào)（如 2）回車(chē)，然后讓你核對(duì)是否選擇對(duì)了編譯器等等，你輸入y回車(chē)。就安裝成功了

>> mex -setup

Please choose your compiler for building external interface (MEX) files: 

Would you like mex to locate installed compilers [y]/n? y

Select a compiler: 

[1] Lcc-win32 C 2.4.1 in C:\PROGRA~1\MATLAB\R2010a\sys\lcc 

[2] Microsoft Visual C++ 6.0 in C:\Program Files\Microsoft Visual Studio 

[0] None 

Compiler: 2

Please verify your choices: 

Compiler: Microsoft Visual C++ 6.0 

Location: C:\Program Files\Microsoft Visual Studio 

Are these correct [y]/n? y

Trying to update options file: C:\Users\Administrator\AppData\Roaming\MathWorks\MATLAB\R2010a\mexopts.bat 

From template: C:\PROGRA~1\MATLAB\R2010a\bin\win32\mexopts\msvc60opts.bat 

Done . . . 

************************************************************************** 

  Warning: The MATLAB C and Fortran API has changed to support MATLAB 

           variables with more than 2^32-1 elements. In the near future 

           you will be required to update your code to utilize the new 

           API. You can find more information about this at: 

           http://www.mathworks.com/support/solutions/en/data/1-5C27B9/?solution=1-5C27B9 

           Building with the -largeArrayDims option enables the new API. 

************************************************************************** 

MATLAB的emd的工具箱  可是不知道怎么用  也不知道怎么用它來(lái)處理txt中的數(shù)據(jù)

安裝中的問(wèn)題

    但是安裝的時(shí)候，如果使用的是VS的編譯器（mbuild –setup、mex –setup設(shè)置），會(huì)報(bào)找不到complex.h的問(wèn)題（用Linux下的Matlab不會(huì)出錯(cuò)），從而使cemdc2_fix.c等文件編譯失敗，這幾個(gè)文件是為了快速實(shí)現(xiàn)計(jì)算EMD而用c編寫(xiě)的，所以即使編譯失敗，也不影響直接使用emd.m實(shí)現(xiàn)EMD功能。如果想編譯成功，可如下修改：（摘自http://www.chinavib.com/thread-79866-1-1.html）

    G. Rilling 07年3月份的程序，運(yùn)行作者的install_emd.m，出現(xiàn)找不到complex.h的問(wèn)題，以下是個(gè)人的理解和解決過(guò)程：（個(gè)人的運(yùn)行環(huán)境為matlab6.5）

complex.h的問(wèn)題

    產(chǎn)生原因：采用matlab的C編譯函數(shù)mex時(shí)，定義了C99_OK的宏（EMDS/make_emdc.m (28行)），利用的是ANSI C99標(biāo)準(zhǔn)如果個(gè)人的電腦中沒(méi)有相關(guān)的支持，就會(huì)出現(xiàn)這個(gè)問(wèn)題。

解決方法：EMDS/make_emdc.m中第28行中mex(’-DC99_OK‘,args(:))語(yǔ)句中的 '-DC99_OK' 即可。

    注意：

    改完之后，運(yùn)行install_emd，會(huì)出現(xiàn)M_PI沒(méi)有定義的問(wèn)題，缺少了常數(shù)PI的宏定義，導(dǎo)致一些.c文件編譯失敗。

    產(chǎn)生原因：去掉C99_OK之后，程序中使用的是作者提供的 emd_complex.h和emd_complex.c兩個(gè)文件來(lái)支持復(fù)數(shù)運(yùn)算，這兩個(gè)文件中，并沒(méi)有定義M_PI這個(gè)宏。

    解決方法：M_PI這個(gè)宏，只在兩個(gè)文件中（clocal_mean.c和clocal_mean2.c）使用，個(gè)人的解決方法是，在相應(yīng)的頭文件（clocal_mean.h和clocal_mean2.h）中加入M_PI的宏定義即可。

    在兩個(gè).h文件中分別加入一下語(yǔ)句：

#define CLOCAL_MEAN_H

#ifndef M_PI

#define M_PI 3.1415926

#endif

    安裝完成后，編譯輸出的.dll文件會(huì)出現(xiàn)，重復(fù)后綴名的問(wèn)題，及 xxx.dll 變成了 xxx.dll.dll自己去掉多余的.dll即可

    最后，關(guān)于版本問(wèn)題：作者推薦使用7.1+版本，但只是針對(duì)個(gè)別的函數(shù)有影響，主要是作者提供的例子程序，無(wú)法在matlab6.5環(huán)境中運(yùn)行，算法的主要功能函數(shù)并不受影響。

工具箱的使用

工具箱函數(shù)

   運(yùn)行help index_emd可以查看工具箱提供的函數(shù)，如下

index_emd.M list of functions in the EMD package
 
 type help function_name for more information on a specific function
 
 Empirical Mode Decomposition
 
   emd         - computes EMD and bivariate/complex EMD with various options
   emd_local   - computes local EMD variation
   emd_online  - computes on-line EMD variation. Note that it does not truly
                  apply on-line: the function is only a demonstration.
   emdc        - fast implementation for EMD with Cauchy-like stopping criterion
                  (requires compilation, see make_emdc function)
   emdc_fix    - fast implementation for EMD with predefined number of iterations
                  (requires compilation, see make_emdc function)
   cemdc       - fast implementation for bivariate/complex EMD (first algorithm)
                  with Cauchy-like stopping criterion (requires compilation,
                  see make_emdc function)
   cemdc_fix   - fast implementation for bivariate/complex EMD (first algorithm)
                  with predefined number of iterations (requires compilation,
                  see make_emdc function)
   cemdc2      - fast implementation for bivariate/complex EMD (second algorithm)
                  with Cauchy-like stopping criterion (requires compilation,
                  see make_emdc function)
   cemdc2_fix  - fast implementation for bivariate/complex EMD (second algorithm)
                  with predefined number of iterations (requires compilation,
                  see make_emdc function)
 
 Utilities
 
   install_emd  - setup Matlab's path and compile the C codes.
   uninstall_emd - revert the modifications made by install_emd and remove the
                   files (optional).
   make_emdc    - compile all C codes
   emd_visu     - visualization of EMD
   cemd_visu    - visualization of bivariate/complex EMD (automatically called
                   by emd_visu when the input is complex)
   cenvelope    - compute envelope curves for bivariate/complex EMD
   cemd_disp    - visualization of envelope curves and tube envelope
   plot3c       - plot a complex vector in 3 dimensions
   plotc        - plot the projection of a complex vector on a variable direction
   dirstretch   - directional stretching of a complex vector
   hhspectrum   - compute Hilbert-Huang spectrum (need the Time-Frequency Toolbox
                   http://tftb.nongnu.org)
   toimage      - transform a spectrum made of 1D functions (e.g., output of
                   "hhspectrum") in an 2D image
   disp_hhs     - display the image output of "toimage" as a Hilbert-Huang spectrum
   addtag       - add a tag to a graphic object (uses the Tag property as a list
                   of keywords or "tags")
   rmtag        - remove a tag from a graphic object (uses the Tag property as
                   a list of keywords or "tags")
   hastag       - test whether a graphic object has a specific tag (uses the Tag
                   property as a list of keywords or "tags")
   findtag      - find objects having a specific tag (uses the Tag property as
                   a list of keywords or "tags")
   
 Examples from G. Rilling, P. Flandrin and P. Gon鏰lves,
   "On Empirical Mode Decomposition and its algorithms"
   IEEE-EURASIP Workshop on Nonlinear Signal and Image Processing
   NSIP-03, Grado (I), June 2003
   
   emd_fmsin        - Fig. 1: a 3-component example (need the Time-Frequency
                       Toolbox http://tftb.nongnu.org)
   emd_triang       - Fig. 2: another 3-component example
   emd_sampling     - Fig. 3: effect of sampling on 1 tone
   emd_separation   - Fig. 4: separation of 2 tones
   ex_online        - Sect 3.4: the way emd_online.m works
   triangular_signal - subroutine called by emd_triang (formerly triang.m)
   
 Examples from G. Rilling, P. Flandrin, P. Gon鏰lves and J. M. Lilly,
   "Bivariate Empirical Mode Decomposition",
   Signal Processing Letters (submitted)
 
   bivariate_EMD_principle       - Fig. 1: principle of the bivariate/complex EMD
   bivariate_EMD_mean_definitions - Fig. 2: definition of the mean for each algorithm.
                                    Also allows to test other signals and parameter sets.
   bivariate_EMD_illustration    - Fig. 3: illustration of the bivariate EMD
                                    on an oceanographic float position record

   稍做整理如下：

EMD分解函數(shù)

工具函數(shù)

來(lái)自《On Empirical Mode Decomposition and its algorithms》的Examples

來(lái)自《Bivariate Empirical Mode Decomposition》的Examples

工具箱使用示例

EMD

clc

clear all

close all

% 原始數(shù)據(jù)

fs = 1000;

ts = 1/fs;

t=0:ts:0.3;

z=2*sin(2*pi*10*t) + 5.*sin(2*pi*100*t);

figure

plot(t, z)

title('原始信號(hào)')

% EMD

imf=emd(z);

emd_visu(z,t,imf)

[A,f,tt]=hhspectrum(imf);

[im,tt]=toimage(A,f);

disp_hhs(im);

邊際譜

clc

clear all

close all

% 原始數(shù)據(jù)

fs = 1000;

ts = 1/fs;

t=0:ts:0.3;

y=2*sin(2*pi*10*t);% + 5.*sin(2*pi*100*t);

figure

plot(t, y)

title('原始信號(hào)')

% 求Hilbert-Huang譜

[A,fh,th] = hhspectrum(y);

figure

subplot(211)

plot(th*ts, A)

title('瞬時(shí)幅值') % 就是包絡(luò)

subplot(212)

plot(th*ts, fh*fs)

title('瞬時(shí)頻率')

% 顯示結(jié)果

[im,tt,ff] = toimage(A,fh,th);

disp_hhs(im,tt)

colormap(flipud(gray))

% 編程實(shí)現(xiàn)顯示

figure

imagesc(tt*ts,[0,0.5*fs],im);

ylabel('frequency/Hz')

set(gca,'YDir','normal')

xlabel('time/s')

title('Hilbert-Huang spectrum')

例子程序

更詳細(xì)的使用說(shuō)明可以參見(jiàn)例子程序，如emd_fmsin.m程序，運(yùn)行結(jié)果如下

EMD分解如下

可以看到，EMD實(shí)現(xiàn)的3個(gè)組分的分離（即分別分解到了IMF1~3中），可見(jiàn)EMD的強(qiáng)大功能。