![](/img/trans.png)
[英]How to loop through matrix elements in mex c++ function for MATLAB?
[英]MEX: How to return a matrix from C++/ C to MATLAB
我可以返回的矩陣只有第一個元素正確,其他所有元素均為零。 我嘗試按照此鏈接中的建議進行操作; 但是,它不適合我。 一些幫助,將不勝感激。
輸出:
>> Matlab_mex_Testing
B1 =
1.0000 0 0
0 1.0000 0
0 0 1.0000
0 0 0
B =
1 0 0
0 0 0
0 0 0
0 0 0
Matlab代碼:
Dir2 = '/home/dkumar/Mex_Codes_DKU/MexCode_Working/Mex_CPP_N_ARMADILLO_Codes_DKU_makefile_Working';
% MEX
cd(Dir2);
[y, B] = normpdfDKU(1/2,0,1);
%Matlab declared
B1 = eye(4,3)
% Identical matrix returned from MEX
B
CPP:
#include "mex.h"
#include <math.h>
#include "/home/dkumar/armadillo-4.600.3/include/armadillo"
using namespace arma;
using namespace std;
#define pi (3.141592653589793)
extern void _main();
const int numInputArgs = 3;
const int numOutputArgs = 2;
// Function declarations.
// -----------------------------------------------------------------
double getMatlabScalar (const mxArray* ptr);
double& createMatlabScalar (mxArray*& ptr);
int TestingLibraries() ; // declared and defined in In Include_4_TSNNLS.c and Include_4_TSNNLS.h
// Function definitions.
// -----------------------------------------------------------------
void mexFunction (int nlhs, mxArray *plhs[],
int nrhs, const mxArray *prhs[]) {
// Check to see if we have the correct number of input and output
// arguments.
if (nrhs != numInputArgs)
mexErrMsgTxt(" DKU-1: Incorrect number of input arguments");
if (nlhs != numOutputArgs)
mexErrMsgTxt("DKU-2: Incorrect number of output arguments");
// Get the inputs.
double x = getMatlabScalar(prhs[0]);
double mu = getMatlabScalar(prhs[1]);
double v = getMatlabScalar(prhs[2]);
// Create the output. It is also a double-precision scalar.
double& p = createMatlabScalar(plhs[0]);
// Compute the value of the univariate Normal at x.
p = exp(-(x-mu)*(x-mu)/(2*v)) / sqrt(2*pi*v);
// CREATE ARMA::mat and print
mat B = eye<mat>(4,3);
//Print B
B.print();
// Trying to return B as second output
mwSize sz[2];
sz[0] = B.n_rows ; // Matlab is row first
sz[1] = B.n_cols ;
//mxArray* pOUT = mxCreateNumericArray(2, sz, mxDOUBLE_CLASS, mxREAL);
plhs[1] = mxCreateNumericArray(2, sz, mxDOUBLE_CLASS, mxREAL);
//Get a pointer to pOUT
double* p2 = (double*)mxGetData(plhs[1]);
*p2 = B[0];
}
double getMatlabScalar (const mxArray* ptr) {
// Make sure the input argument is a scalar in double-precision.
if (!mxIsDouble(ptr) || mxGetNumberOfElements(ptr) != 1)
mexErrMsgTxt("The input argument must be a double-precision scalar");
return *mxGetPr(ptr);
}
double& createMatlabScalar (mxArray*& ptr) {
ptr = mxCreateDoubleMatrix(1,1,mxREAL);
return *mxGetPr(ptr);
}
罪魁禍首就在這里:
*p2 = B[0];
這等效於僅復制B
的第一個值並將其放入p2
的第一個插槽中,該插槽是輸出的第一行,第一列。 您需要復制整個矩陣。
還要記住,MATLAB中的矩陣是列專業的,因此您需要了解索引到p2
。 但是, Armadillo指出,如果僅使用單個線性索引訪問矩陣,則您將以列優先格式訪問元素 。 因此,而不是做*p2 = B[0];
, 做這個:
for (int i = 0; i < B.n_elem; i++)
p2[i] = B[i];
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.