tutorial
visualization

Introducing Text Annotations in Foxglove’s Image Panel

Annotate your robots’ images with metadata for easier debugging

tutorial
visualization

Foxglove users can now annotate their camera images with the TextAnnotation message schema – whether you’re using ROS, Protobuf, or any other supported serialization format.

Text annotations can dramatically enrich how you debug your robots’ perception modules. By rendering useful metadata in-context, text annotations can facilitate faster comprehension and richer analysis.

bounding boxes

Labeling bounding boxes can help roboticists quickly understand how their perception stack is performing.

To display your own annotated images in Foxglove, check out our text-annotation-demo in the MCAP GitHub repo. This simple example renders a ball bouncing around a scene over time, and publishes two TextAnnotation messages to label the scene with the timestamp (top left corner of the scene) and the ball’s position (above our perceived bouncing box).

Run the demo script to output a text-annotation-example.mcap file for visualization in Foxglove. Add an Image panel to your layout, and toggle on the “annotations” topic in the panel settings to display both labels:

panel settings

While this exercise demonstrates a rather simple use case for annotated images, we hope it inspires your team to annotate your images – with perception labels, error messages, or other useful metadata.

For more information on image annotations, you can check out our docs or contact us directly in our Discord community with any questions.

Read more

Start building with Foxglove.

Get started for free