camera coordinate system of visualization

[yanghtr]

I want to know the camera coordinate definition of capture_depth_float_buffer. I search the document and issue and find no place to mention it. By drawing the world coordinate system frame, I know that the world coordinate system is x right, y up, -z forward (-z points to inner).
When I use the function capture_depth_float_buffer to capture depth, the depth is captured by a camera posed using the current view point. However, what is the camera coordinate? Is it a x right, y down, z forward camera? Or some other camera coordinate system, such as x right, y up, z forward?

[theNded]

In computer vision, the conventional coordinate system is x right, y down, z forward.

    ^
   /
 z
/
--- x --->
|
y
|
v

We consistently follow this convention in Open3D.
OpenGL has a different convention, but we have handled the conversion in capture_depth_buffer. As long as you do not access this internal representation, you don't have to worry about it.