I'm trying to convert depth data to pixels(8 bit).
I tried the default formula:
BYTE intensity = static_cast<BYTE>((depth >= nMinDepth) && (depth <= nMaxDepth) ? (depth % 256) : 0);
pRGBX->rgbRed = intensity;
pRGBX->rgbGreen = intensity;
pRGBX->rgbBlue = intensity;
However that has a lot of banding issues. I need to get rid of the banding.
So what I did was take the min & max depth values from the kinect, and use them to normalize any depth value the kinect sends me.
After that I would just multiply the normalized value with 255.
Result:
The banding is gone however there are still issues such as: there's not much diversity in range, i mean the values are either very close to black, or quite close to white. And most importantly, the artifacts, especially those in the upper right corner.
Any ideas for improvements?
