Multi Camera Depth Image stitching for point cloud generation

Hello. I’m trying to generate a stitched pointcloud from depth images obtained via multiple camera. The cameras have different focal length (one longer range camera with higher focal length and small fov, and two large fov short range cameras). I wanted to ask two questions on how to use openCV for this:

A) Is it better to generate pointcloud for individual images and then stich them or stich the depth images to create a panorama, and then generate a pointcloud?

B) which algos to use for either case.