By simulating the image and depth seen in the camera coordinated and having the camera motions, it is possible to add typical distortions and get high fidelity rgb-d outputs.
The following distortions are modeled:
Temporal (time based) distortions including: Pure delay, Exposure time(motion blur), Rolling shutter, and Interlacing.
Geometric (Lens) distortions: 5parameter Brown-Conrady lens distortion
Photometric distortions: Daylight, Dynamic and Static pixel-wise noise, and Vignetting.
Installation angle errors
for more information, see the simulator code at:
https://github.com/mah92/camsim
It is written with c++/glsl