Using the python openCV interface, I did this by calling cvConvertScale:
cvConvertscale(32bitImage, 8bitimage, 256)
where the 32bit image is the thing I want to convert; the 8 bit
(unsigned int) image is the target/destination image (which might be
what you refer to as a frame buffer), and 256 indicates the scale to
linearly adjust to. The documentation all shows 1.0 instead of 256 -
this makes a black image for me (or more likely, a binary image that
looks black because it's all 0's and 1's on a scale of 256).
> Is there a way of doing this conversion directly? I was able to do this by
> cvConvertScale'ing to an intermediate 8-bit grayscale and then
> cvCvtColor'ing to 3 channels.
> Can I also do this without the intermediate frame buffer?