This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

opencv Dsp aceleration

Dear All,

I'am trying to use the DSP on the beagleBoard. I'am doing it based on this wiki: http://code.google.com/p/opencv-dsp-acceleration/

So far I was able to use the sobel algorithm and the rgbToGRay(with some modifications), running on DSP.

I'am addig algoritmos based on these tutorials:

--> http://processors.wiki.ti.com/index.php/C6Accel_Advanced_Users_Guide

--> http://code.google.com/p/opencv-dsp-acceleration/wiki/Procedure_To_add_other_OpenCV_Algorithms

The problem is that I was not able to use the original color conversion APIS. These APIs are available here:

--> http://code.google.com/p/opencv-dsp-acceleration/source/browse/trunk/dsp_opencv/beagle/library/opencv/beagle_opencv_API.c#224

--> http://code.google.com/p/opencv-dsp-acceleration/source/browse/trunk/dsp_opencv/beagle/library/opencv/beagle_opencv_API.c#258

These APIs uses as argument, in the main application,  pointer the the output from the cvQueryFrame.

--> http://code.google.com/p/opencv-dsp-acceleration/source/browse/trunk/dsp_opencv/beagle/apps/opencv/linux/app.c#332

I was not able to make this APIs work. The function UNIVERSAL_processAsync, form the c6accelw.c, returns an error. 

I did my own rgbToGray algorithm  implementation, based on the  DFT API that is available here:

-->  http://code.google.com/p/opencv-dsp-acceleration/source/browse/trunk/dsp_opencv/beagle/library/opencv/beagle_opencv_API.c#82

And the input arguments were also based on the DFT example that you can see here:

-->http://code.google.com/p/opencv-dsp-acceleration/source/browse/trunk/dsp_opencv/beagle/apps/opencv/linux/app.c#582

When i use the cvAlloc function to allocates memory and puts the video data on this allocation all the process works fine.

The problem is that I'am always have to copy the video information to the allocated memory. Is there a way to do it directly, passing the the output from the cvQueryFrame directly?

Why i can't run the originals color convertions algoritms?

Best Regards,

Carlos