Ričards's profile - activity

Thanks for replies. Will try to rebuild with Qt support. Weird thing is that previously building OpenCV I didn't have to explicitly set GTK+ support. Also CMake GUI version doesn't seem to include that option.

Hello.

I Built opencv 3.0 master branch from source using cmake and visual studio 2013 community version. Aditional support was enabled for CUDA and VTK. Everything built fine, but when I call function imshow(), I get following error in console:

opencv error: Unspecified error (function ins not implemented. Rebuild the library with windows, GTK+ 2.x or Carbon support)....

It seems that it has something to do with window UI framework. In cmake configuration Win32UI option was selected and no errors appeared while building

thanks.

Fix: By following suggestions in comments I rebuild OpenCV with Qt support (beforehand installing Qt 5.5 itself) and it solved the problem.

Hello.

I need to output Mat data to csv file. I remember doing it with function format(matrix, "csv"). That formats matrix data to string using csv format. But now using OpenCV 3.0 alfa build I get error for this function in Visual Studio, saying something about wrong parameters. Is this functionality gone from OpenCV?

Thank you :)

I am experiencing this error when receiving frames from IP-camera:

[mpeg4 @ 0x147c7e0] Context scratch buffers could not be allocated due to unknown size.

Weird thing is that this error doesn't appear every time I run the program. Any suggestions? Thenks

Hello.

I am experiencing problems connecting to two IP-Cameras simultanously using VideoCapture object. For some time I am able to receive frames normally, but after some time (about few seconds), the camera was connected first, starts to return empty frames. Also this error message is displayed [mjpeg @ 0x7fe888008c00] overread 8. If I connect only one camera, then everything works fine.

My steps are as follows:

//Declare video capture objects for camera streams:

VideoCapture streamCam1; VideoCapture streamCam2;

// open camera streams for specified ip addre; streamCam1.open(address1); streamCam2.open(address2);

// loop for capturing frames from camera, and visualizing them while(((waitKey(1))&0xFF)!=27){ streamCam1>>image1; streamCam2>>image2; imshow("CAM1", image1); imshow("CAM2", image2); }

Thank You.

Hello.

I have problems finding how to change the descriptor size of ORB feature detector/descriptor. Looks like default setting is 32 bytes, but I think there should be possibility to change that.

Also I would like to know if patchSize in ORB() parameters is size of the one side of the patch, or is it the pixel area of the patch?

Thanks