Description
Indoor robots without GPS rely on cameras, and CPU visual SLAM eats all the compute. Isaac ROS Visual SLAM uses NVIDIA-accelerated cuVSLAM for real-time visual odometry, mapping and localization.
It supports stereo, multi-camera and IMU fusion at high frame rates even on Jetson, as a ROS 2 package.
Odometry:Real-time pose.
SLAM:Loop closure.
Multi-camera:Stereo and more.
IMU fusion:More robust.
It supports stereo, multi-camera and IMU fusion at high frame rates even on Jetson, as a ROS 2 package.
Features
Odometry:Real-time pose.
SLAM:Loop closure.
Multi-camera:Stereo and more.
IMU fusion:More robust.
