Brixus's profile - activity

Thank you :-)

Hi,

I would like to use my webcam as a photo camera.

I would like to have a function like "Mat GetOneFrame();" which I can call in short distances to deliver just one up to date frame.

This is my try:

void TakeOneFrame()
{
    namedWindow("Video");
    namedWindow("Picture");
    Mat frame;
    Mat picture;

    VideoCapture cap (0);
    for(int i=0; i<=2;i++)
       {
        if(char(waitKey(1)) != 'q' &&cap.isOpened())
            {cap.operator >>(frame);}
        frame.copyTo(picture);
        imshow("Video", frame);
       }
    cap.release();


   while(1)
    {
        imshow("Picture",picture);
    }
}

I found out, that I need to grab a few frames until there is one sharp one. The Video window produces the desired frame, but when the window "Picture" opens up, there is no image showing. Even if I chance the position of frame.copyTo(picture); this makes no change.

Why is that and is there a better solution for my problem?

Thank you very much :-)

Thanks, this works, but is to much work ;-) I just needed to convert my image as FooBar stated and then I used drawContours on that :-)

Thank you very much. :-) Sometimes one does not see all the trees because of the forest ;-)

Hello,

I have a grayscale image (CV_8UC1) and I use findContours to identify contours. These contours I would like to draw colored into the grayscale image.

I tried to merge two empty Mats(CV_8UC1) together with the grayscale image and this works not really well. So I have my wished colors, but the original image is blue, red or green depending on the merged channel position.

How can I achieve the grayscale image staying gray with having colored contoures on top?

Thank you very much :-)

Hi,

I just installed OpenCV on a RaspberryPi.

I tried the tutorial given here: http://docs.opencv.org/doc/tutorials/introduction/linux_gcc_cmake/linux_gcc_cmake.html#linux-gcc-usage

Futhermore I compiled this code also with QMake and added some console outputs.

I get no errors and the console outputs are working, but in general there shows no picuture or even a frame up when I run the executable file.

Do you have any idea why it is like this and how to solve it?

Thank you very much :-)

No, I am using Ubuntu.

The error is: "terminate called after throwing an instance of 'std::logic_error' what(): basic_string::_S_construct null not valid"

If it is really a linker problem, how can I solve it? I use OpenCV a lot with the C++ functions and there are no other errors.

Hello,

I would like to use the function "createTrackbar" for creating a trackbar on the QT control panel.

The info about this feature you can find here: http://docs.opencv.org/modules/highgui/doc/qt_new_functions.html

If I use "cvCreateTrackbar" with NULL as window everything works fine but with createTrackbar it ist not working.

Has anybody an idea how to fix this?

This is a working minimal:

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

using namespace cv;    

const int slider_1_max  = 255;
      int slider_1 = 0;
const int slider_2_max  = 255;
      int slider_2 = 0;

Mat test;


void on_trackbar_1(int pos)
{
pos=slider_1;    
}

void on_trackbar_2(int pos, void*)
{    
}


int main ()
{
    namedWindow("Test");

     cvCreateTrackbar("One",NULL, &slider_1,slider_1_max, on_trackbar_1);

     //Not working:
     //createTrackbar("Two",NULL, &slider_2,slider_2_max, on_trackbar_2);


 while (char(waitKey(1)) != 'q'){}
 return 0;
}

Thank you very much :-)

Thank you very much :-)

Hello,

I use the function "convexHull" and I would like to know how I can access every vector point of a convex hull object.

For example I use this code:

              /// Find the convex hull object for each contour  
                 vector<vector<Point> >hull( contours_vec.size() );
                 for( int i = 0; i < contours_vec.size(); i++ )
                    {  convexHull( Mat(contours_vec[i]), hull[i], false ); }

              /// Draw contours + hull results Mat
                 Mat drawing = Mat::zeros( Dilated.size(), CV_8UC3 );  
                 for( int i = 0; i< contours_vec.size(); i++ )
                    {
                      Scalar color = Scalar( rng.uniform(0, 255), rng.uniform(0,255), rng.uniform(0,255) );
                      drawContours( drawing, contours_vec, i, color, 1, 8, vector<Vec4i>(), 0, Point() );
                      drawContours( drawing, hull, i, color, 1, 8, vector<Vec4i>(), 0, Point() );

If I now use "cout << hull[i] << endl;" I get something like this for every contour:

[812,176; 810,182; 805,190; 812, 173].

How can I access just for example the point 805, 190?

I tried hull[i].something, but it did not work.

Thank you very much. :-)

Thank you very much. This sounds logical :-)

Hi everybody, I would like to blend pictures manually using a loop. In detail I want to copy the white pixels from a binary image (0==black, 255==white, not other values possible) to a new Picture which is at the beginning a Mat just including black pixels.

My code works fine if I use a Mat created with white pixels. But if I use my binary images I load into OpenCV I get artefacts. I cannot understand why this is happening. If I look in the source image in OpenCV the pixels have just 0 or 255, but if I vary with the thresholdvalue I get different Artefacts.

Normaly intensity >=1 should work well, but If I use for example 100 I get still differend artefacts.

This is the code I use:

#include <opencv2/highgui/highgui.hpp>
#include <iostream>

using namespace std;
using namespace cv;

void Vblend(Mat src, Mat blend)
 {   int colu, row;
     int intensity;
     for (row=0; row<=480; row++)
     {
         for (colu=0; colu<=640; colu++)
         {
             intensity = src.at<uchar>(row, colu);
             if(intensity >=1) {blend.at<uchar>(row, colu)=255;}
         }
     }
 }


int main()
{
   Mat src= imread("pic.jpg",CV_LOAD_IMAGE_GRAYSCALE);
   Mat blend(480,640,CV_8UC1,Scalar(0));

    namedWindow("src");
    namedWindow("blend");

   Vblend(src,blend);

   imshow("src", src);
   imshow("blend",blend);
   while(char(waitKey(1)) != 'q') {}
   return 0;
}

In this picture you can see the original on the left and the artefacts on the right.

blend.PNG

How can I eliminate the artefacts?

Thank you very much for your help :-)

Hello,

I try to get familiar with OpenCV. Therefore I am playing with the Camshift demo included in the documentation.

I would like to improve the code thats if the tracked object leaves the window I dont get an error. The goal is that a tracked object leaves the window and that the tracking stopps until the object is present again in the window.

Sadly I don't habe any idea how to solve this as I am a beginner in OpenCV.

At the moment I get the following error if the object leaves the window:

"Bad Argument (Input window has non-positive sizes) in cvMeanShift, camshift.cpp, line 8."

Has anybody an idea how to solve this problem?

Thank you very much :-)

Hello, I would like to convert a stream from my webcam to HSV colorspace.

Therefore I used cvtColor in the following code:

int main()
{
VideoCapture cap(0);
if(!cap.isOpened())
{
cout << "Capture could not be opened succesfully" << endl;
return -1;
}
namedWindow("RGB");
namedWindow("HSV");

while(char(waitKey(1)) != 'q' && cap.isOpened())
{
Mat frame, frame_hsv;
cap >> frame;
cvtColor(frame, frame_hsv, CV_RGB2HSV);

imshow("RGB", frame);
imshow("HSV", frame_hsv);

}
return 0;
}

First I am wondering because if I change CV_RGB2HSV to CV_BGR2HSV there is no difference in the shown images.

In the attached file you can see the output.

Why are the colors in the new "HSV-image" wrong? And what can I do to solve this?

I thougt cvtColor would calculate the matching colors for the new color space.

Thank you :-)

RGBtoHSV.PNG

Thank you very much.

Your answer gave me the right hint.

Qt5 was not installed correct. On Ubuntu it is not enough to use the Qt installer.

This guide here solved my problems: Guide

Ater this I could install OpenCv without problems. :)

Thank you very very much :-) I forgot to tell that I use Linux, my fault. Sorry for this :-/

Hello,

I am not able to install OpenCV 2.4.6.1 properly. Poorly Cmake does not find the QT stuff and I get the following warning:

"CMake Warning at cmake/OpenCVFindLibsGUI.cmake:20 (find_package): By not providing "FindQt5Gui.cmake" in CMAKE_MODULE_PATH this project has asked CMake to find a package configuration file provided by "Qt5Gui", but CMake did not find one.

Could not find a package configuration file provided by "Qt5Gui" with any of the following names:

Qt5GuiConfig.cmake
qt5gui-config.cmake"

I know the correct path to this things, but I do not know how to tell it Cmake. Also with google I found no solution.

This is the normal command I use: cmake -D WITH_TBB=ON -D BUILD_NEW_PYTHON_SUPPORT=ON -D WITH_V4L=ON -D INSTALL_C_EXAMPLES=ON -D BUILD_EXAMPLES=ON -D WITH_QT=ON -D WITH_OPENGL=ON ..

How can I add the path to the Qt sources or is there another solution?

Thank you very much :)

Thanks for the answers. I just saw them now. I switched to linux ;-)

Thanks, but in the linker options for debugging everything has got the "d", but it is still not working. I guess its a problem with .dll files, but I dont know.

He just says Programpath\Progname.exe cannot be started. Unknown Mistake in Framework of Windows-Webservices"

Maybe the first part of the error which I forgot to mention earlier os also important ;) Thanks!

Thank you very much for your answer.

I tried a lot of different guides and ended up with this guide: http://4someonehelp.blogspot.in/2013/06/solved-install-opencv-245-using-visual.html I replaced 245 with 246 in the .lib files to make it compatible.

Your suggestion was quiet good. I changed the dir and OpenCV is still creating working .exe files, but in Debug-Mode I get still the same error as mentioned above.

Hi everyone,

I am a complete beginner with programming and OpenCV. I managed to install OpenCV 2.4.6. on Win 8 x64 with Visual Studio 11 and it took me round about 10 hours ;)

If I tried to compile a project in ReleaseMode I could not start the .exe because of a missing .dll. But the correct path is added to the systems variables! Than I just copied the dlls to the Windows Systems folder and now the release .exe works.

But if I try to run OpenCV-Code in VS in Debug Mode I get an error. Freely translated it just says:"Unknown Mistake in Framework of Windows-Webservices".

What can I do to solve this problem?

Sorry for my bad english and thank you very much :-)