Embodied AI Glossary中文

Meshcat

Advanced

A lightweight visualization tool that displays robots and 3D scenes in a web browser.

Meshcat is an open-source 3D visualization tool Robin Deits developed during his PhD at MIT, built on WebGL (the browser's 3D graphics interface) and three.js. A backend program sends meshes, point clouds, coordinate frames, and poses over WebSocket to a web page, which can then be opened in a browser to view and rotate — no desktop graphics environment needed, which makes it well suited to remote servers and Jupyter notebooks. Drake has it built in, and libraries such as Pinocchio provide a Meshcat interface too; it's commonly used to check inverse-kinematics results, replay trajectories, or inspect collision geometry.

ExampleIn Jupyter, load a URDF with Pinocchio, call MeshcatVisualizer, and watch the arm's pose update in the browser as joint angles change.

Also called
meshcat-python
Related
Drake · Pinocchio · Viser · RViz / RViz2 · Unified Robot Description Format (URDF)
Sources
meshcat GitHub
meshcat-python GitHub

See it in the full glossary →