This tutorial shows how to efficiently record Kinect data into the ROS bag file format, and play it back into the same ROS processing graph.

Begin here. This tutorial demonstrates how to open a Kinect (or other OpenNI device) in ROS, introduces ROS tools for visualization and configuration, and explains how to get registered (depth + RGB) outputs like color point clouds.

This tutorial shows how to calibrate the intrinsic parameters of the IR (depth) and RGB cameras, including focal length and the distortion model. For applications requiring high accuracy, calibration can improve on the default camera models used by openni_camera. This calibration does not necessary fully calibrate the depth. Especially the ASUS Xtion is prone to depth errors and needs further calibration (see also here and here). There is no official module in ROS yet to calibrate this (as of 6th February 2015).

This tutorial demonstrates how to register the Kinect depth image to an external RGB camera instead of the built-in one. This requires an extrinsic calibration, calculating the relative transform between the two cameras.

This tutorial demonstrates an alternative to OpenNI for registering the depth and RGB cameras based on a user calibration instead of the factory one. It requires an extrinsic calibration, calculating the relative transform between the two cameras.