Open bmegli opened 4 years ago
There are several ways to optimize but first we should try to speed up the alignment.
From librealsense#5128 we see that CPU alignment itself at larger resolutions is time consuming.
I have no plans to use CUDA now(no NVIDIA on my Realsense host).
There was some work in librealsense for GLSL processing blocks.
There is also GLSL example.
From #22 for color aligned depth + color with L515 we clearly don't keep up with the data rate