Skip to content

Conversation

@heweweh
Copy link

@heweweh heweweh commented Nov 24, 2022

I'm trying to generate training data with carla-simulator, and fetch image/pointcloud/objects information via carla-ros-bridge.
the problem is the message at /carla/ego_vehicle/objects or /carla/objects topic with [derived_object_msgs](https://docs.ros.org/en/kinetic/api/derived_object_msgs/html/index-msg.html)/Object type.

I found the object_sensor is publishing objects like following behaviors:

  1. if the object is a vehicle, then the Object.pose.position will be the position on the ground(x-y plane), but it will be adjust to pivot point when publishing Marker.
  2. if the object is a walker, then the Object.pose.position will be the pivot point of bounding box, and there is no adjustment before publishing Marker.

the reason, that I consider these behaviors are problematic, is the behaviors are not uniform, so I made this patch.
the reason, that I choose to modify the walker.py, is that the position is intentionally adjusted when publishing Marker in vehicle.py, so I think maintainers are proposing to publish objects topic with position on the ground, but not the pivot point.

please help me to confirm this MR, thanks.


This change is Reviewable

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant