Hello, I tested Isaac ROS VSLAM in Isaac Sim with Nova Carter and saw that the performance could be improved if it could use the depth image also from the 3D camera.
In the documentation of PyCuVSLAM it is specified that it can use the depth image from the RGB-D camera for tracking. Can Isaac ROS VSLAM be configured for this tracking as well?
Please let me know what could be done.
Hello @stefan133,
Welcome to the Isaac ROS forum and thanks for your post!
Isaac ROS VSLAM does not officially expose a “depth-aided tracking” mode, even though the underlying library supports it. However, you can run VSLAM on stereo input and use the RGB-D camera’s depth image with NVBlox or another mapper, feeding in the VSLAM pose as the camera transform like this. This setup can improve your environment model and collision avoidance, even if tracking itself is stereo-only.