Transforming pose of Node from one anchor to another anchor #277
Replies: 1 comment
-
Only the parent nodes participate in calculation of the world transform. If there is a parent
This is a caching logic since calculating the world transform can be expensive if there are many parent nodes.
Yes, that is a simple mathematical process of multiplying the transform matrices recursively as you can see in the following method. Again, the camera transform isn't required to compute the world transform of the node. |
Beta Was this translation helpful? Give feedback.
-
I want to compute a new pose for the node to position it on a new anchor. My initial idea was to use the pose of the first anchor and the second anchor to bring the node from the reference frame of the first anchor to second anchor:
Then I noticed that sceneform provides world position and rotation of the node as well. So, I decided to use those:
Now, my question is, how does sceneform compute the pose of a simple node in the world coordinate frame? Does it use the pose of the camera and anchors to recompute the pose of the node? I tried following the logic here but there are a lot of keywords that don't really make sense (WORLD_POSITION_DIRTY, LOCAL_TRANSFORM_DIRTY etc.). Is the process of recomputing the pose of the node in the world coordinate system a simple mathematical process (using the latest and previous estimated camera pose provided by ARCore to update the position of all Nodes or something along that line) or is there some computer vision stuff being used as well?
Beta Was this translation helpful? Give feedback.
All reactions