using OpenCV FileStorage in LabVIEW

[SilvaShi]

Excuse me,

I read almost all the vis and cpp files, and I find one image out, it's a kinect'RGB image.

But I still haven't found how to return an Mat to the LabVIEW and display it.

Could you please tell me which project is or how to do it?

Thank you.

[Klemen]

Hello,

every example using OpenCV consists of passing the image (actually the image array) pointer to Mat type. Take a look at the Hough circle transformation for example. The basic idea is:

Mat inputImage(rows,cols,CV_8U,&imdata[0]);

where imdata is the pointer to your image array.

Best regards,

K

[SilvaShi]

Thank you.

Yes, actually I have already known how to get an image from the LabVIEW and transfer it to the OpenCV Mat before,

But I still don't know how the OpenCV dll return the image to the LabVIEW.

Can the dll return the RGB image, or can only return the CV_8U image and need to combine the three channels in LabVIEW?

Best Regards

[Klemen · Accepted solution]

Hello,

something like this should work:

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

#include <opencv2/core/core.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <opencv2/highgui/highgui.hpp>

using namespace std;
using namespace cv;

class myclass{
public:    
    Mat img;
    int readimage(){
        img = imread("C:\\Users\\Public\\Pictures\\Sample Pictures\\Desert.jpg", CV_LOAD_IMAGE_COLOR);
        if(!img.data)
            return -1;
        return 0;        
    }
};

myclass temp;

extern "C"{
    _declspec (dllexport) int readimage(int *img_rows, int *img_cols);
    _declspec (dllexport) int image_processing(uchar *data);
}

_declspec (dllexport) int readimage(int *img_rows, int *img_cols)
{
    int error_code = temp.readimage();
    if(error_code != 0)
        return -1;
    (*img_rows) = temp.img.rows;
    (*img_cols) = temp.img.cols;
    return 0;
}

_declspec (dllexport) int image_processing(uchar *data)
{
    Mat img_to_LV(temp.img.rows, temp.img.cols, CV_8U, &data[0]);
    cvtColor(temp.img, img_to_LV, CV_BGR2GRAY);
    return 0;    
}

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

and:

You can also pass the color image, but take care as OpenCV uses bgr and not rgb notation. In the past I have used 3 seperate channels, building them back in Labview.

Best regards,

K

[SilvaShi]

I did it!

Thank you so much, klemen. It takes you a lot of time.

Thank you for your attention!