matlab opencv混合编程（“mex找不到指定模块”）

使用mex 可以将c/cpp程序变成matlab的接口函数，提升效率，通用性的配置方式这里不再累述，主要介绍使用opencv库的混合编程

安装opencv至某个目录，将其运行库（动态库dll）添加到环境变量中

>D:\opencv\build\x64\vc10\bin;

64位系统添加使用x64,32位系统使用x86如果不清楚，或者会在vs中有使用opencv的32位项目，可以添加都添加

>D:\opencv\build\x64\vc10\bin;D:\opencv\build\x86\vc10\bin

在matlab下mex -setup选择编译器

编写好包含OpenCV库的Matlab与C/C++混合编程代码

#include 
#include 

#include 
#include 

#include "mex.h"

// Matlab entry point function
void mexFunction( int nlhs, mxArray *plhs[], int nrhs, const mxArray*prhs[] )
{
    // Check if the input argument is legal
    if ( nrhs != 1 || !mxIsChar( prhs[0] ) )
    {
        mexErrMsgTxt("An image name should be given.\n");
    }

    // Get the name of the image
    int nStringLen;
    nStringLen = mxGetNumberOfElements(prhs[0]);
    std::string szImageName;
    szImageName.resize( nStringLen + 1 );

    mxGetString( prhs[0], &szImageName[0], nStringLen + 1 );

    // Read the image from file
    cv::Mat image;
    image = cv::imread( szImageName );

    // Show the image if it is successfully read from disk
    if ( !image.empty() )
    {
        cv::imshow( "Test Mex OpenCV", image );
    }
    else
    {
        mexErrMsgTxt("The specified image does not exist.\n");
    }
}

使用mex编译 需要制定opencv头文件、静态库文件的目录，已经要链接的静态库的名称，注意opencv的版本号

mex OpenCVShowImage.cpp -ID:\opencv\build\include 
-LD:\opencv\build\x64\vc10\lib -lopencv_core243 -lopencv_imgproc243 -lopencv_highgui243

其中-IF:\3rdlibs\OpenCV\include，告诉编译器可以在D:\opencv\build\include这个目录进行头文件的查找；
-LD:\opencv\build\x64\vc10\lib，告诉链接器可以在D:\opencv\build\x64\vc10\lib这个目录进行库文件的查找；
最后三个参数表面我们写的源代码需要链接opencv_core243、opencv_imgproc243和opencv_highgui243这三个OpenCV的静态库。

将opencv/build/x86/vc10/bin的路径加到matlab的路径中

这是非常关键的一步

Matlab关于这个问题其实是有官方说明的

If MATLAB® cannot find all .dll files referenced by a MEX file, it
cannot load the MEX file. MATLAB displays the following error message:

Invalid MEX-file : The specified module could not be
found. where mexfilename is the module with the dependency error. This
module cannot find its dependent libraries. To resolve this error,
find the names of the dependent libraries, and determine if they are
present on your system and on the system path.

On Windows® systems, to find library dependencies, use the third-party
product Dependency Walker. Dependency Walker is a free utility that
scans any 32-bit or 64-bit Windows module and builds a hierarchical
tree diagram of all dependent modules. For each module found, it lists
all the functions exported by that module, and which of those
functions are called by other modules. Download the Dependency Walker
utility from the website http://www.dependencywalker.com/. See
http://www.mathworks.com/matlabcentral/answers/92362-how-do-i-determine-which-libraries-my-mex-file-or-stand-alone-application-requires
for information on using the Dependency Walker.

For .dll files that the MEX file linked against when it was built, the
.dll files must be on the system path or in the same folder as the MEX
file.

MEX files might require additional libraries that are not linked to
the MEX file. Failure to find one of these explicitly loaded libraries
might not prevent a MEX file from loading, but prevents it from
working correctly. The search path used to find these explicitly
loaded libraries is controlled by the code that loads the libraries
and might not include the folder that contains the MEX file. Consult
the library documentation on proper installation locations.

Possible reasons for failure include:

MATLAB version incompatibility Missing compiler runtime libraries. If
your system does not have the same compiler that built the MEX file,
see the Microsoft® MSDN® website for information about Visual C++®
Redistributable Packages. Missing or incorrectly installed specialized
runtime libraries. Contact your MEX file or library vendor.