Jarosław Karwowski

@IntoSan, you simply have to define a `camera_optical` frame so `tf` can find such a frame in a transform tree. You can find an example in a related package: [diff_drive_mapping_robot](https://github.com/rayvburn/diff_drive_mapping_robot/blob/6fd5ba0c73210272faa0da22685ee86fe064b7d7/diff_drive_mapping_robot/src/tf_broadcaster.cpp#L59)....

Sure that Kinetic is pretty old but e.g. PAL's software for [`TiAGo robot`](https://pal-robotics.com/robots/tiago/) seems to work best with that distro. Having that in Kinetic would be very helpful!

Hello! Not really sure what you are planning to do with the point cloud - some more details of the application context could be helpful. However, if I understand correctly,...

@MagdaZal this is true, however I will keep my reply in English so English-speakers could also benefit from this conversation. So, the main question is - is your camera monocular?...

So, having a monocular camera, you should also feed my modification of ORB-SLAM2 with the odometry data to perform scale correction. The resultant point cloud should be pretty accurate. Have...