Administrator
|
We can use RGB-D camera + 3D lidar for SLAM. However, only camera or lidar will be used to create the 3D point cloud. To combine point cloud from camera and lidar, you would have to merge them prior to rtabmap, then feed that point cloud to rtabmap instead of the 3d lidar.
For odometry, we woudl use either rgbd_odometry or icp_odometry using one or the sensor. There is also a node called rgbdicp_odometry, which would use Reg/Strategy=2 to compute odometry, thus computing a first guess using vision, then refine that pose with lidar.
cheers,
Mathieu
|