Hi,
Have you tried to build the scene graph online? I mean use a real-time odometry and RGB-D camera to provide sensor data and odometry when the robot is moving, and setup a process to take them as input to build the scene graph.
This is theoretically possible, but we didn't try that.
The merging process is a bit slow and can not run in real-time, you are also bound by how fast SAM+CLIP inference time
Hi, Have you tried to build the scene graph online? I mean use a real-time odometry and RGB-D camera to provide sensor data and odometry when the robot is moving, and setup a process to take them as input to build the scene graph.