Openni: spatial coordinate system (original)

Joint positions and orientations are given in the real world coordinate system. the origin of the system is at the sensor. + X points to the right of the, + y points up, and + z points in the direction of increasing depth. the coordinate frame is shown in the figure above.
The joint position and direction are based on the real-world coordinate system. The coordinate origin is located in sensor. The positive direction of the X axis is the right side of the human body facing the sensor. The positive direction of the Y axis is upward. The positive direction of the Z axis is the direction in which the depth value is increased. (This is a typical right-hand coordinate system)

Joint positions are measured in units of mm.
The joint position is measured in millimeters.

Joint orientations are given as a 3x3 rotation (orthonormal) matrix. This represents a rotation between the joint's local coordinates and the world coordinates.
The first column is the direction of the joint's + X axis given as a 3-vector in the world coordinate system.
The second column is the + Y axis ction, and the third column is the + z axis direction.
Our "neutral pose" is the T-pose shown in the figure above. in this pose, each joint's orientation is aligned with the world coordinate system. that is, its orientation is the identity matrix.
The orientation of the joint is based on a 3x3 rotation (orthogonal) matrix. This matrix depicts the rotation between the local coordinates of the joint and the Global coordinates.
The first column of the Matrix represents the positive direction of the X axis of the joint, a 3-vector in the world coordinate system.
The second column of the Matrix represents the positive direction of the Y axis of the joint, and the third column of the Matrix represents the positive direction of the Z axis of the joint.
Our "neutral pose" is a T-pose, as shown in. In this posture, the direction of each joint is consistent with the world coordinate system. That is to say, its direction is the unit matrix.
 

Known issues known issues

1.Arm tracking is less stable when the arm is close to other body parts, especially the torso. if both arms are close to the torso, as well as to each other, they might get mixed up.
When the arm is close to other parts of the body (especially the trunk), the stability of Arm tracking is poor. If the arms are very close to the trunk and the other side, they may be confused.

2.Leg tracking is still somewhat unstable and noisy. It works better when the user stands with legs separated. Fast motions and complex kicks or crouches might cause the tracking to fail.
Leg tracking has always been unstable and interfering. When you separate your legs, the tracking effect will be better. Quick movements and complex kicks or squats may cause tracing failures.

3.Pose tracking may also become somewhat unstable if the head is not visible.
If the header is invisible, the posture tracking will become unstable.

4.Arms and legs in extremely stretched positions (I. e. Near the limits of human flexibility) might be lost by the tracker.
When the arm and leg are stretched to the limit (for example, reaching the limit of human flexibility), tracking may be lost.

5.If the skeleton is stuck in a faulty pose, or stuck facing the opposite ction, then returning to a "simple" pose (arms away from torso so the sensor can see them and legs separated) shocould help resolve it.
If the skeleton gets stuck in a wrong posture or faces the opposite direction, return to a simple posture (for example, moving the arm away from the trunk and standing separately on both legs .) These will help solve the problem of getting stuck.

6.In general, very fast motions may cause tracking failure.
In general, a fast motion may cause a tracing failure.

7.In some cases, overall tracking might be bad. Re-calibrating the user may resolve the problem.
In some cases, the overall tracking effect may be poor. Then "calibrating" will solve the problem.