Hi Hector,

On Tuesday 11 March 2008, Héctor Hugo Avilés Arriaga wrote:
> Dear all,
>
> I'm a newbie in all this webcam stuff and trying the Luvcview source code.
> Luvcview is useful to me because it allows to display images from two
> videcameras at the same time. Doing the same with OpenCV retrieves
> corrupted images.
>
> I want to access the pixel information. I'm able to convert
> YUYV format to 24RGB easily using the Pyuv422torgb24 function,
> and the display the images using a 24bits BGR format directly.

You should contact Michel Xhaard for questions related to luvcview internals. 
He is the luvcview author.

> Unfortunately, I cannot do the same with the MJPEG format.
> I realize that the image is decompressed into the
> videoIn->framebuffer buffer before displaying it, but I don't know how to
> access the pixel information correctly.
> I have tried to copy the image into an OpenCV buffer in this way:
>
> opencv->imageData[curr_pos] = videoIn->framebuffer[curr_pos++]; // B
> opencv->imageData[curr_pos] = videoIn->framebuffer[curr_pos++]; // G
> opencv->imageData[curr_pos] = videoIn->framebuffer[curr_pos++]; // R
>
> However, all I see are vertical lines and
> my silhouette repeated three times on the image.
> As you can see,  I'm assuming this information is stored in BGR format
> in the videoIn->framebuffer.
>
> The questions are: which is the correct form to access the information
> captured with the MJPEG format?  and, using this format speeds up the
> capturing process after all?

I'm not familiar with the luvcview code, but the data might be stored after 
decompression in YUV format instead of RGB. Just a wild guess.

Best regards,

Laurent Pinchart
_______________________________________________
Linux-uvc-devel mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/linux-uvc-devel

Reply via email to