Best way to get stereo camera depth data from FPGA -> RPi ??

[technix]

If you can emulate the signal given by the few MIPI CSI2 devices that the Pi supports, you can use that interface to send an uncompressed video feed to the Pi, and try to pack your depth map data into the pixels.

This is one place where Pi has its shortcomings: it isn’t easy to add support for CSI and DSI devices. If it were Allwinner V3s (using the recently mainlined 8-bit camera interface) things can be a bit easier.