ROS Resources: Documentation | Support | Discussion Forum | Service Status | Q&A

Vision AI Standard Interfaces

Hi all, friendly neighborhood navigator here.

I wanted to bring up the migration of vision_msgs so graciously donated by @Kukanani ( from his personal GitHub to ros-perception to aid in our use of vision_msgs as the standardized AI Detection and Classification interfaces in both 2D and 3D.

vision_msgs have been used for some time in various projects but until now it wasn’t “official” and not broadly adopted (projects define their own sets of messages to translate the same sets of information). The aim of bringing vision_msgs into the ROS organization is to standardize around this package so that all future AI packages can be easily swapped out (as is enabled by sensor_msgs, geometry_msgs, etc.).

If you have a vested interest in AI detection / classification interfaces, please check out the link below and lets chat if there’s any changes we think are appropriate to adopt these at a community level.

It’s a great pleasure to be able to pull in some more wonderful work from the community. I want to again thank @Kukanani for his great work on this package and willingness to work with us to create this community standard.

If you have some wonderful work you’d like pulled into main ROS, feel free to drop me a PM.

Happy detecting,