[Vxl-users] vidl2 convert to vil_image_view

[Junaed S. <ju...@ci...>]

Hi again.
So I am still working on using vidl2 for image sequences. I can read a
sequence successfully, but when I try to convert the next frame
(vidl2_frame) to a vil_image_view<vxl_byte>, I get some unexpected
results. My images are all RGB color images, with dimensions of 510
pixels by 422 pixels (just an example), 3 bytes per pixel (RGB). I can
read them individually using vil_load. I have this code snippet:

    vil_image_view<vxl_byte> image;
    vidl2_frame_sptr frame = inputStream.current_frame();
    if( !vidl2_convert_to_view( frame, image ) )
    {
        return false;
    }

I hope I'm doing the right thing to convert a vidl2_frame to
vil_image_view. If so, then here's the problem:
istep (next pixel. row wise) should be at every 3 bytes, and jstep (next
row, same column) should be at width*nplanes bytes. But it isn't. With
the numbers above, the values of istep should be 3 and jstep should be
1566(3*522) pixels, which is what it is if loaded using vil_load. In the
case above with vidl2_convert_to_view, the values are istep=1,
jstep=width (522), in the object 'image'.

If the above code is wrong, I'd love to know the right way of doing
things. Any pointers would be greatly appreciated! Thanks.

-J-


PS: Debugger shows this when using vil_load: istep_ = 3, jstep_ = 1566,
planestep_ = 1.
      Debugger shows this when using vidl2_convert: istep_ = 1, jstep_ =
522, planestep_ = 214020