which image type of Depth should I choose of RGBD recording

[brz-hub]

Which data type is the depth data of rgbd camera, DepthPlanar, DepthPerspective or DepthVis? In ToF principle, I think we should choose DepthPerspective, but many people choose DepthVis or DepthPlanar. So which type should I choose to record data?

[CloudsRL]

Hello, i faced the same trouble, have you solve it? i need help too

[CloudsRL]

3ku so much, i will try it. I am tring to use AirSim to make dataset and use it in my SLAM program, but this problem confuse me very long time. I try to make depth into pointcloud and then turn it into Octomap, but i falled QAQ i hope this will help.

[brz-hub]

I hope you succeed in your program! :)

[koyukan]

3ku so much, i will try it. I am tring to use AirSim to make dataset and use it in my SLAM program, but this problem confuse me very long time. I try to make depth into pointcloud and then turn it into Octomap, but i falled QAQ i hope this will help.

How was it? I am facing a similar challenge :D @CloudsRL

[CloudsRL]

thanks @brz-hub , i have already finished it ^^
DepthVis is not good, because it is for look.
DepthPerspective is ok, but i don't use it. Because when i use it, my uav feel the environment like a ball, so weird.
DepthPlanar is good, i use it. And by the way, i set the depth in pfm format, and then i use python to change it into the format i want. In this condition, u could setting the solution and distance as u like.(from my python, the longest distance is about 600m? but from my point, the RGB-D could only feel about 10-20m, so i think that's enough^^)

[CloudsRL]

@koyukan i m not good at GitHub, i hope u could see what i write down for you