Output image depth from airsim to achieve obstacle aviodance


I want to output airsim’s depth information to Ardupilot in SITL to achieve obstacle avoidance target. Everything is just in simulation. Ardupilot has already have the computer-vision based obstacle avoidance algorithm. This is the algorithm I am talking about where a ZED stereo camera is used.


So I want to know how to output the airsim’s depth to SITL to achieve this target. Could you please give some instructions? Thank you.